Aider
v0.86.0 open-sourceaider --model xai/grok-4
aider --commit-language Japanese
aider --add-gitignore-files --model gemini-2.5-pro
aider --analytics-posthog-host https://posthog.internal.example.com --analytics-posthog-project-api-key <your-key>
/settings
aider --attribute-co-authored-by --no-attribute-author --no-attribute-committer
aider --model grok3
aider --model quasar
# While in an aider session, press Ctrl-X Ctrl-E to open the current input buffer in $EDITOR
/context add authentication middleware to validate JWT tokens on all API routes
aider --git-commit-verify true
AWS_PROFILE=my-security-profile aider --model bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0
aider --check-model-accepts-settings false --reasoning-effort high --model openrouter/google/gemma-3-27b-it
/think-tokens 8k
aider --architect --auto-accept-architect <file>
aider --notifications --notifications-command 'notify-send "Aider" "Response ready"'
aider --model openrouter/deepseek/deepseek-chat:free
aider --model sonnet
# In your .aider.model.settings.yml:
- name: deepseek/deepseek-r1
use_temperature: 0.2
aider --model o3-mini --reasoning-effort low
aider --model openrouter/deepseek/deepseek-r1:free
aider --model r1
aider --model openrouter/deepseek/deepseek-r1
/ask
# ... ask questions ...
/code
# ... request edits ...
/architect
aider --model openrouter/deepseek/deepseek-chat
# ... AI? Why does this function return None instead of raising an exception?
aider --multiline
/copy-context Focus on security vulnerabilities in the authentication flow
aider --copy-paste
aider --api-key anthropic=sk-ant-<your-key>
aider --watch-files
aider --voice-input-device 2
aider --timeout 120
aider --no-detect-urls
aider --show-release-notes=false <file>
# In ~/.aider.model.settings.yml
- name: aider/extra_params
extra_params:
temperature: 0.2
max_tokens: 4096
/web https://example.com/api-reference
aider --haiku
aider --apply-clipboard-edits file-to-edit.js
aider> /save context.aider
# later or in another session:
aider> /load context.aider
aider --load project_context.aider --model gpt-4o
aider --no-fancy-input
/read-only src/**/*.py
aider --skip-sanity-check-repo --model gpt-4o
aider --voice-format mp3
aider --o1-mini
aider --model o1-preview
/report
aider --chat-language Spanish
aider --no-suggest-shell-commands
/settings
aider --suggest-shell-commands
aider --upgrade
aider --install-main-branch
/reset
aider --map-multiplier-no-files 4
aider --map-refresh files
aider --deepseek
aider --chat-mode ask
# then inside the session:
/code refactor the authentication module to use bcrypt
aider --read ../shared/schema.sql src/models.py
/clipboard
aider --message "/read ../policy.md Summarize any security policy violations in auth.py" auth.py
cd services/auth && aider --subtree-only
aider --map-tokens 1024 --model <your-model>
/add-clipboard-image
aider --commit-prompt 'Write a commit message prefixed with the Jira ticket number from the branch name, followed by a short imperative summary.'
docker run --user $(id -u):$(id -g) -v $(pwd):/app paulgauthier/aider-full
aider
/ask What does the authentication flow in auth.py do?
# .aider.conf.yml
file: CONVENTIONS.md
/chat-mode help
aider --just-check-updated
/help How do I set a custom model settings file?
aider --attribute-commit-message
aider --model-settings-file ./my-model-settings.yml
aider --sonnet
aider --sonnet --llm-history-file llm-messages.log
aider --vim
{
"my-custom-model": {
"max_tokens": 8192,
"input_cost_per_token": 0.000002,
"output_cost_per_token": 0.000006
}
}
aider --lint-cmd 'flake8 --max-line-length 120' mymodule.py
aider --test-cmd 'pytest tests/' --test mymodule.py
aider --restore-chat-history
aider --show-prompts
aider --browser
aider --models gemini
aider --model groq/llama3-70b-8192 --no-show-model-warnings
aider --opus --weak-model claude-3-haiku-20240307
aider --sonnet --anthropic-api-key <your-api-key>
aider --model gpt-4-turbo
aider -3 <file>
aider -4 <file>
aider --openai-organization-id org-xxxxxxxxxxxxxxxx
/web https://nvd.nist.gov/vuln/detail/CVE-2024-12345
/test pytest tests/
aider --openrouter --model mistralai/mistral-7b-instruct
! pytest tests/
aider --4-turbo
aider --commit
/run (cd subdir; ls)
aider --voice-language fr
aider --no-stream
/git log --oneline -5
/git log --oneline -10
aider --show-repo-map
aider --dark-mode
/add src/**/*.py
aider --encoding latin-1
aider --encoding latin-1 src/legacy_module.py
/add src/**/*.py
aider --code-theme monokai main.py
/add src/**/*.py
aider --encoding latin-1 myfile.c
aider --encoding cp1252 myfile.py
aider --code-theme monokai myfile.py
aider --message 'Refactor all print statements to use logging' src/app.py
aider --no-stream src/app.py
aider --openai-api-base http://localhost:8080/v1 --model gpt-3.5-turbo-16k
aider --map-tokens 4096 --model gpt-4-32k
/tokens Summary
Aider is an open-source command-line tool that lets a developer pair-program with an LLM directly in the terminal, editing an existing codebase or scaffolding a new one and committing the results to git as it goes. It connects to a wide range of cloud and local models, including current Claude, GPT-5, Gemini, and DeepSeek releases, so a developer can switch models mid-session rather than being locked to one provider. It builds a map of the repository to give the model context on large codebases, and it is aimed at individual developers and small teams who want AI-assisted changes tracked through normal git commits rather than pasted from a chat window. First committed in 2023, it has drawn 191 contributors and 108 commits in the past year, with its own release notes noting that Aider now writes most of its own new code.
What Aider answers
Which languages does the repository map understand beyond the mainstream ones?
it tags Fortran, Haskell, Julia, Zig, OCaml, MATLAB, and Clojure among others, so context-aware editing extends to less common codebases, not just JavaScript and Python
Do I have to touch configuration every time a new model comes out?
model aliases like sonnet, opus, gemini, and flash are kept pointed at the current release, so a session can move to a newer model without editing settings
What happens to my commit history when the AI makes the change?
each change lands as a normal git commit, with an option to add a co-author trailer identifying the AI's contribution
Can it work with files my repository is set up to ignore?
yes, files normally excluded by git can be explicitly added to its editing scope when needed
Is this still an actively maintained tool or a stalled project?
the most recent tracked release was 377 days ago against a history of 108 commits in the past year, so activity has slowed relative to its earlier pace
Will it keep working if I'm still on an older Python setup?
no, support for Python 3.9 has been dropped, so upgrading the runtime is required before adopting a current release
Examples
Command line
No option matches that search.
| option | found in | since | description |
|---|
No option matches that search.
Values are placeholders taken from each option’s declared default. Nothing is executed here — the output shown is a recording of a run that already happened.
Release history
- v0.86.0
Aider v0.86.0 adds GPT-5, Grok-4, Kimi-K2, and Gemini Flash Lite support with improved
/clearand/undoUX.└──▷ GET THIS VERSION$ git clone --branch v0.86.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.86.0
└──▷ TRY ITSwitch to Grok-4 mid-session to leverage its capabilities on a task.$ aider --model xai/grok-4
- ›Supports all GPT-5 models.
- ›Supports Grok-4 via
xai/grok-4andopenrouter/x-ai/grok-4model names. - ›Supports
gemini/gemini-2.5-flash-lite-preview-06-17model. - ›Supports
openrouter/moonshotai/kimi-k2model. - ›The
/clearcommand now prints "All chat history cleared." as confirmation.
+2 moreshow less
- ›The
/undocommand now shows only the first line of each commit message for easier reading. - ›Running
/modelwith no arguments now displays model announcements.
- v0.85.0
Aider v0.85.0 adds Gemini 2.5 & o3-pro support, commit language control, and gitignore-file editing.
└──▷ GET THIS VERSION$ git clone --branch v0.85.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.85.0
└──▷ TRY ITGenerate commit messages in a non-English language for localized team workflows.$ aider --commit-language Japanese
Include files normally excluded by .gitignore (e.g., build artifacts or generated configs) in Aider's editing scope.$ aider --add-gitignore-files --model gemini-2.5-pro
Route analytics to a self-hosted PostHog instance instead of the default endpoint.$ aider --analytics-posthog-host https://posthog.internal.example.com --analytics-posthog-project-api-key <your-key>
- ›Adds support for
gemini-2.5-pro,gemini-2.5-flash, andgemini-2.5-pro-preview-06-05with thinking tokens;flashandgeminialiases updated accordingly. - ›Adds support for OpenAI o3-pro and o1-pro via the Responses API across multiple providers.
- ›Enables disabling thinking tokens by setting them to
0, with improved help text and examples. - ›New
--add-gitignore-filesflag allows adding .gitignore-listed files to Aider's editing scope. - ›New
--commit-languageoption lets users specify the language for generated commit messages.
+6 moreshow less
- ›Co-authored-by attribution is now enabled by default in commit messages.
- ›New
--analytics-posthog-hostand--analytics-posthog-project-api-keyflags enable custom PostHog analytics configuration. - ›Adds MATLAB language support for repository maps.
- ›Adds Clojure language support for repository maps.
- ›Increases max tokens for Deepseek models to 65,536.
- ›Skips expensive file tracking operations when
--skip-sanity-check-repois set, improving performance.
- ›Adds support for
- v0.84.0
Aider v0.84.0 adds Claude Sonnet 4/Opus 4 support, shell tab completion, and auto-refresh for GitHub Copilot tokens.
└──▷ GET THIS VERSION$ git clone --branch v0.84.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.84.0
└──▷ TRY ITInspect which model variants (main, editor, weak) are active and their capabilities before starting a session.$ /settings- ›Supports new Claude Sonnet 4 and Opus 4 models (
claude-sonnet-4-20250514,claude-opus-4-20250514); defaultsonnetandopusaliases updated. - ›Supports
vertex_ai/gemini-2.5-flash-preview-05-20model. - ›Updates default OpenRouter onboarding models to
deepseek/deepseek-r1:free(free tier) andanthropic/claude-sonnet-4(paid tier). - ›Automatically refreshes GitHub Copilot tokens when used as OpenAI API keys.
- ›Adds shell tab completion for file path arguments and
--edit-format/--editor-edit-formatoptions.
+2 moreshow less
- ›The
/settingscommand now displays detailed metadata for active main, editor, and weak models. - ›Introduces a local cache for OpenRouter model metadata, increasing reliability and performance.
- ›Supports new Claude Sonnet 4 and Opus 4 models (
- v0.83.0
Aider v0.83.0 adds shell completions, Playwright scraping, co-author commits, OCaml repo-map, and new model support.
└──▷ GET THIS VERSION$ git clone --branch v0.83.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.83.0
└──▷ TRY ITAttribute AI-assisted commits with a co-author trailer while suppressing author/committer rewriting.$ aider --attribute-co-authored-by --no-attribute-author --no-attribute-committer
- ›Adds support for
gemini-2.5-pro-preview-05-06andqwen3-235bmodels. - ›Adds repo-map support for OCaml and OCaml interface files.
- ›Introduces
--attribute-co-authored-byflag to add a co-author trailer to commit messages, with--attribute-author/--attribute-committeroverrides for fine-grained control. - ›Adds
--disable-playwrightflag to prevent Playwright installation prompts and usage. - ›Adds
--shell-completionsargument to generate shell completion scripts (bash, zsh, etc.).
+9 moreshow less
- ›Enables
aider scrapeCLI tool to use Playwright for web scraping when available. - ›Automatically fetches model parameters (context window, pricing) for OpenRouter models from their website.
- ›Enables
thinking_tokensandreasoning_effortparameters for OpenRouter models. - ›Enables
reasoning_effortfor Gemini 2.5 Flash models. - ›Tracks total tokens sent and received, now included in benchmark statistics.
- ›Displays token count progress and file/identifier name during repo map updates.
- ›The
aider-argsutility now defaults to printing a sample YAML configuration when run with no arguments. - ›Improves
/askmode to instruct the LLM to elide unchanging code in responses. - ›Commit message prompt now specifies the user's language.
└──▷ BREAKING ON UPGRADE- !Dropped support for Python 3.9; upgrading will break setups running on Python 3.9.
- ›Adds support for
- v0.82.0
Aider v0.82.0 adds GPT-4.1/grok-3/Gemini 2.5 Pro support, new patch and editor edit formats, and Fireworks AI deepseek-v3.
└──▷ GET THIS VERSION$ git clone --branch v0.82.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.82.0
└──▷ TRY ITUse the grok3 alias to quickly target xai/grok-3-beta without typing the full model path.$ aider --model grok3
- ›Supports GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano models.
- ›Improved architect mode support for Gemini 2.5 Pro.
- ›Adds support for
xai/grok-3-beta,xai/grok-3-mini-beta,grok-3-fast-beta,grok-3-mini-fast-beta, and OpenRouter variants includingopenrouter/openrouter/optimus-alpha. - ›New
patchedit format for OpenAI's GPT-4.1 model. - ›New
editor-diff,editor-whole, andeditor-diff-fencededit formats.
+3 moreshow less
- ›Adds short aliases:
grok3forxai/grok-3-betaandoptimusforopenrouter/openrouter/optimus-alpha. - ›Allows adding files by full path even when a file with the same basename is already in the chat.
- ›Adds support for Fireworks AI model
deepseek-v3-0324.
- v0.81.0
Aider v0.81.0 adds Quasar Alpha model support and OpenRouter OAuth authentication flow.
└──▷ GET THIS VERSION$ git clone --branch v0.81.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.81.0
└──▷ TRY ITUse the free Quasar Alpha model on OpenRouter for AI-assisted coding without any API cost.$ aider --model quasar
- ›Adds support for
openrouter/openrouter/quasar-alphamodel (currently free on OpenRouter), accessible viaaider --model quasar. - ›Offers OpenRouter OAuth authentication automatically when an OpenRouter model is specified but no API key is present.
- ›Adds model metadata for
openrouter/google/gemini-2.0-flash-exp:free. - ›Configures
gemini/gemini-2.0-flashandopenrouter/google/gemini-2.0-flash-exp:freeas weak models for Gemini 2.5 Pro models.
- ›Adds support for
- v0.80.0
Aider v0.80.0 adds OpenRouter OAuth, Ctrl-X Ctrl-E editor keybinding, Scala repomap support, and smarter model defaults.
└──▷ GET THIS VERSION$ git clone --branch v0.80.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.80.0
└──▷ TRY ITEdit a long, complex prompt in your preferred terminal editor mid-session instead of typing it inline.$ # While in an aider session, press Ctrl-X Ctrl-E to open the current input buffer in $EDITOR- ›Adds OpenRouter OAuth flow: prompts to authenticate via OAuth when no model or API keys are provided, and auto-selects a default OpenRouter model based on free/paid tier status when
OPENROUTER_API_KEYis set. - ›Prioritizes
gemini/gemini-2.5-pro-exp-03-25whenGEMINI_API_KEYis set, andvertex_ai/gemini-2.5-pro-exp-03-25whenVERTEXAI_PROJECTis set, as the default model. - ›Adds Ctrl-X Ctrl-E keybinding to open the current input buffer in an external editor.
- ›Adds repomap support for the Scala language.
- ›Boosts repomap ranking for files whose path components match identifiers mentioned in chat, improving context relevance.
+6 moreshow less
- ›Validates user-configured color settings on startup, warning and disabling invalid ones.
- ›Warns at startup when
--streamand--cache-promptsare used together, since cost estimates may be inaccurate in that combination. - ›Changes web scraping timeout from a hard error to a warning, allowing scraping to continue with partial content.
- ›Left-aligns markdown headings in terminal output for improved readability.
- ›Updates edit format to the new model's default when switching models with
/model, if the user was using the previous model's default format. - ›Adds the
openrouter/deepseek-chat-v3-0324:freemodel.
- ›Adds OpenRouter OAuth flow: prompts to authenticate via OAuth when no model or API keys are provided, and auto-selects a default OpenRouter model based on free/paid tier status when
- v0.79.0
Aider v0.79.0 adds Gemini 2.5 Pro and DeepSeek V3 support plus a new
/contextcommand for automatic file scoping.└──▷ GET THIS VERSION$ git clone --branch v0.79.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.79.0
└──▷ TRY ITLet Aider automatically determine which files are relevant before making a change, instead of manually adding them to context.$ /context add authentication middleware to validate JWT tokens on all API routes- ›Adds support for SOTA Gemini 2.5 Pro model.
- ›Adds support for DeepSeek V3 0324 model.
- ›New
/contextcommand automatically identifies which files need to be edited for a given request. - ›Adds
/editas an alias for the/editorcommand. - ›New "overeager" mode for Claude 3.7 Sonnet models to constrain it to the requested scope.
- v0.78.0
Aider v0.78.0 adds thinking-token support for OpenRouter Sonnet 3.7, new model-switching commands, AWS profile auth for Bedrock, and git hook control.
└──▷ GET THIS VERSION$ git clone --branch v0.78.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.78.0
└──▷ TRY ITPreserve your repo's pre-commit hooks (e.g. linters, secret scanners) so Aider commits pass through them normally.$ aider --git-commit-verify true
Use a named AWS profile for Bedrock authentication without hardcoding credentials.$ AWS_PROFILE=my-security-profile aider --model bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0
Force reasoning-effort settings onto a model that doesn't advertise support, bypassing validation.$ aider --check-model-accepts-settings false --reasoning-effort high --model openrouter/google/gemma-3-27b-it
- ›Adds thinking-token support for OpenRouter Sonnet 3.7 models.
- ›New
/editor-modeland/weak-modelcommands let you switch model roles interactively mid-session. - ›New
--git-commit-verifyflag (default: false) controls whether git commit hooks are bypassed on Aider commits. - ›New
--check-model-accepts-settingsflag (default: true) validates--reasoning-effortand--thinking-tokensagainst model capabilities, ignoring unsupported settings — or force them through by setting the flag to false. - ›Adds
AWS_PROFILEsupport for Bedrock models, enabling named AWS profile auth instead of explicit credentials.
+2 moreshow less
- ›Adds support for the
openrouter/google/gemma-3-27b-itmodel. - ›Improves code block rendering in markdown output with better padding via NoInsetMarkdown.
- v0.77.0
Aider v0.77.0 adds 150 new tree-sitter languages, in-session reasoning controls, and an auto-accept architect flag.
└──▷ GET THIS VERSION$ git clone --branch v0.77.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.77.0
└──▷ TRY ITCap reasoning token spend during a session to control cost when using extended-thinking models.$ /think-tokens 8kRun an architect-mode session that automatically applies proposed changes without manual confirmation prompts.$ aider --architect --auto-accept-architect <file>
- ›Adds support for 130 new languages with linter support and 20 new languages with repo-map support via tree-sitter-language-pack.
- ›New
/think-tokenscommand sets thinking token budget mid-session using human-readable formats (e.g., 8k, 10.5k, 0.5M); displays current setting when called with no arguments. - ›New
/reasoning-effortcommand controls model reasoning level interactively; displays current setting when called with no arguments. - ›New
--auto-accept-architectflag (default: true) skips confirmation prompts when accepting changes from the architect coder format. - ›Adds
ignore_permission_deniedoption to the file watcher to silently skip restricted files instead of erroring.
+3 moreshow less
- ›Adds model support for
cohere_chat/command-a-03-2025andgemini/gemma-3-27b-it. - ›The bare
/dropcommand now preserves original read-only files supplied via--read. - ›
--thinking-tokensargument now accepts human-readable string values (e.g., 8k, 0.5M).
- v0.76.0
Aider v0.76.0 adds thinking-token budgets, desktop notifications, and broader model support including Claude 3.7 Sonnet and GPT-4.5-preview.
└──▷ GET THIS VERSION$ git clone --branch v0.76.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.76.0
└──▷ TRY ITGet a desktop ping when Aider finishes a slow LLM call so you can stay focused on other work.$ aider --notifications --notifications-command 'notify-send "Aider" "Response ready"'
Use the free DeepSeek V3 model via OpenRouter for cost-free AI-assisted coding.$ aider --model openrouter/deepseek/deepseek-chat:free
- ›Adds
--thinking-tokensflag to control the token budget for thinking/reasoning models that support it. - ›Displays thinking/reasoning content returned by LLMs inline in the session.
- ›Adds
--notificationsflag to trigger desktop alerts when an LLM response is ready and awaiting input. - ›Adds
--notifications-commandto specify a custom desktop notification command. - ›Adds support for QWQ 32B model.
+6 moreshow less
- ›Adds support for DeepSeek V3 free tier via OpenRouter (
openrouter/deepseek/deepseek-chat:free). - ›Adds support for Claude 3.7 Sonnet on OpenRouter, Bedrock, and Vertex AI (including
:betavariant). - ›Updates default OpenRouter model to Claude 3.7 Sonnet.
- ›Adds support for GPT-4.5-preview model.
- ›Offers to install dependencies automatically for Bedrock and Vertex AI models.
- ›Switches tree-sitter support to
tree-sitter-language-pack.
└──▷ BREAKING ON UPGRADE- !The
remove_reasoningsetting is deprecated and replaced byreasoning_tag; existing configs usingremove_reasoningwill trigger a deprecation warning. - !Model shortcut args such as
--4oand--opusare deprecated in favor of--model.
- ›Adds
- v0.75.0
Aider v0.75.0 adds Claude 3.7 Sonnet, HCL/Terraform syntax, tree-sitter language pack, and o1/o3-mini markdown generation.
└──▷ GET THIS VERSION$ git clone --branch v0.75.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.75.0
└──▷ TRY ITSwitch to the new Claude 3.7 Sonnet model for an Aider session.$ aider --model sonnet
- ›Adds basic support for Claude 3.7 Sonnet via
--model sonnet. - ›Adds HCL (Terraform) syntax support for infrastructure-as-code workflows.
- ›Adds support for the tree-sitter language pack, expanding language parsing coverage.
- ›Adds
openrouter/o3-mini-highmodel configuration. - ›Adds
build.gradle.ktsas a recognized special file for Kotlin project support.
+1 moreshow less
- ›Enables markdown output for o1 and o3-mini models by sending the 'Formatting re-enabled.' string.
- ›Adds basic support for Claude 3.7 Sonnet via
- v0.74.0
Aider v0.74.0 adds dynamic Ollama context sizing, finer temperature control, and faster startup across more providers.
└──▷ GET THIS VERSION$ git clone --branch v0.74.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.74.0
└──▷ USE ITPin a precise temperature for a model in your model settings file instead of just enabling/disabling it.# In your .aider.model.settings.yml: - name: deepseek/deepseek-r1 use_temperature: 0.2
- ›Dynamically resizes the Ollama context window to fit the current chat automatically.
- ›Adds
use_temperature: <float>support in model settings for precise temperature control, beyond the previous true/false toggle. - ›Improves support for o3-mini, DeepSeek V3, DeepSeek R1, o1-mini, and o1 via secondary API providers.
- ›Strips
<think>tags from DeepSeek R1 responses when used as the weak model (e.g., for commit messages). - ›Full Docker container now bundles
boto3for AWS Bedrock integration.
+5 moreshow less
- ›Docker containers now set
HOME=/appto persist~/.aiderhistory at the standard project mount-point. - ›Watch files now fully skips top-level directories listed in ignore files (e.g.,
node_modules), reducing the chance of hitting OS inotify limits. - ›Faster startup when model metadata is supplied via local files or when using more providers.
- ›Improved
.gitignorehandling: honors ignores already in effect from any configuration source and checks.envonly when the file exists. - ›Yes/No prompts now accept All/Skip as aliases for
Y/Neven outside group-confirmation flows.
- v0.73.0
Aider v0.73.0 adds full o3-mini and DeepSeek R1 support with new
--reasoning-effortcontrol.└──▷ GET THIS VERSION$ git clone --branch v0.73.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.73.0
└──▷ TRY ITRun o3-mini with reduced reasoning verbosity to keep context usage low during large refactors.$ aider --model o3-mini --reasoning-effort low
Use DeepSeek R1 free tier via OpenRouter for cost-free reasoning-model sessions.$ aider --model openrouter/deepseek/deepseek-r1:free
- ›Adds full support for OpenAI o3-mini via
aider --model o3-mini. - ›New
--reasoning-effortflag acceptslow,medium, orhighto tune reasoning model verbosity. - ›New
remove_reasoning: <tagname>model setting strips model-specific reasoning tags from responses. - ›Supports DeepSeek R1 free tier on OpenRouter via
--model openrouter/deepseek/deepseek-r1:free. - ›Auto-creates parent directories when creating new files in a session.
+1 moreshow less
- ›Case-insensitive model name matching while preserving original case.
- ›Adds full support for OpenAI o3-mini via
- v0.72.0
Aider v0.72.0 adds DeepSeek R1 support, Kotlin repo-map parsing, and configurable line endings.
└──▷ GET THIS VERSION$ git clone --branch v0.72.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.72.0
└──▷ TRY ITUse DeepSeek R1 as the coding model for a session — useful when evaluating a reasoning-focused model against your codebase.$ aider --model r1
Use DeepSeek R1 via OpenRouter to route through a managed API endpoint instead of the direct DeepSeek service.$ aider --model openrouter/deepseek/deepseek-r1
- ›Supports DeepSeek R1 via
--model r1shortcut or OpenRouter with--model openrouter/deepseek/deepseek-r1. - ›Adds Kotlin syntax support to the repo map for accurate code context.
- ›New
--line-endingsflag controls line-ending style when writing files. - ›Adds read-only file announcements so practitioners can see which files Aider will not modify.
- ›Supports DeepSeek R1 via
- v0.71.0
Aider v0.71.0 adds custom voice device settings, doubles chat history limit, and makes bare mode-switch commands easier.
└──▷ GET THIS VERSION$ git clone --branch v0.71.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.71.0
└──▷ TRY ITSwitch chat mode on the fly without arguments to quickly pivot between asking questions and requesting code changes.$ /ask # ... ask questions ... /code # ... request edits ... /architect- ›Bare
/ask,/code, and/architectcommands now switch the active chat mode without requiring additional arguments. - ›Increases max chat history tokens from 4k to 8k, enabling longer context for complex sessions.
- ›Increases default repomap size for broader codebase awareness.
- ›Adds support for custom voice format and input device settings.
- ›Adds token count feedback when adding command output to chat.
+3 moreshow less
- ›Streaming automatically disables for models that don't support it, enabling seamless switching between
/model o1and streaming models. - ›Improves markdown rendering performance with adaptive delay based on render time.
- ›Pretty output remains enabled when editing files containing triple-backtick fences.
- ›Bare
- v0.70.0
Aider v0.70.0 adds full o1 model support, OpenRouter DeepSeek integration, git-free operation, and subtree-scoped file watching.
└──▷ GET THIS VERSION$ git clone --branch v0.70.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.70.0
└──▷ TRY ITUse the DeepSeek Chat model via OpenRouter without managing API keys directly.$ aider --model openrouter/deepseek/deepseek-chat
- ›Adds full support for o1 models.
- ›Supports
openrouter/deepseek/deepseek-chatas a selectable model. - ›Enables running Aider without git installed.
- ›Watch files now respects
--subtree-only, limiting file watching to that subtree. - ›New install methods via uv, including one-liner invocations.
+1 moreshow less
- ›Shows hints about
AI!andAI?syntax when user makes AI comments in watched files.
- v0.69.0
Aider v0.69.0 adds watch-all-files AI comments, multiline input mode, Gemini Flash 2.0 Exp support, and richer copy-context.
└──▷ GET THIS VERSION$ git clone --branch v0.69.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.69.0
└──▷ TRY ITTrigger Aider from inside any source file to ask a question without switching to the terminal.$ # ... AI? Why does this function return None instead of raising an exception?Enable multiline input so you can compose multi-paragraph prompts without accidentally sending early.$ aider --multiline
Copy code context with custom instructions to paste into a web UI like ChatGPT or Claude.$ /copy-context Focus on security vulnerabilities in the authentication flow- ›Supports
# ... AI?comments in any file to trigger Aider and ask inline questions about code. - ›Watch mode now monitors all files (not just selected source types), with AI comment support via
# AI,// AI, or-- AIsyntax. - ›Adds
--multilineflag and/multiline-modecommand to make ENTER a soft newline and META-ENTER send the message. - ›Full support for Gemini Flash 2.0 Exp via
--model flashor--model gemini/gemini-2.0-flash-exp. - ›
/copy-contextnow accepts optional instructions to include when copying code context to the clipboard.
+1 moreshow less
- ›
/voicenow lets you edit the transcribed text before sending.
- ›Supports
- v0.68.0
Aider v0.68.0 adds copy-paste mode for LLM web UIs, CLI API key management, and bash/zsh watch-file support.
└──▷ GET THIS VERSION$ git clone --branch v0.68.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.68.0
└──▷ TRY ITUse Aider with a web-based LLM UI (e.g., Claude.ai) when you have no direct API access.$ aider --copy-paste
Supply your Anthropic API key directly on the command line without setting environment variables.$ aider --api-key anthropic=sk-ant-<your-key>
- ›New
--copy-pastemode enables Aider to work with LLM web chat UIs (e.g., ChatGPT, Claude.ai) instead of requiring a direct API connection. - ›New
/copy-contextcommand copies the current context to the clipboard for pasting into an LLM web chat UI. - ›New
--api-key provider=keysetting lets you supply API keys for any provider directly from the command line or a YAML config file. - ›New
--set-env VAR=valuesetting lets you configure arbitrary environment variables for LLM providers from the command line or YAML config. - ›Adds bash and zsh support to
--watch-files, enabling auto-detection of AI coding comments in shell scripts.
+1 moreshow less
- ›Adds experimental Gemini models.
- ›New
- v0.67.0
Aider v0.67.0 adds in-editor AI comment instructions, Amazon Bedrock Nova support, and smarter
/diffand/runworkflows.└──▷ GET THIS VERSION$ git clone --branch v0.67.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.67.0
└──▷ TRY ITRun aider in watch mode so you can drop AI instructions directly in your source files without switching to a terminal.$ aider --watch-files
- ›New
--watch-filesmode lets aider watch source files for inline AI instructions — trigger execution by ending a comment withAI!. - ›Supports new Amazon Bedrock Nova models.
- ›The
/diffcommand now invokesgit diff, honouring your configured diff tool. - ›Adds Ctrl-Z support for process suspension.
- ›New
- v0.66.0
Aider v0.66.0 adds PDF support for Sonnet/Gemini, voice device selection, API timeout control, and improved
/dropmatching.└──▷ GET THIS VERSION$ git clone --branch v0.66.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.66.0
└──▷ TRY ITSelect a specific microphone when your system default isn't the one you want for voice input.$ aider --voice-input-device 2
Set a longer API timeout to avoid premature failures when working with slow or large model responses.$ aider --timeout 120
- ›Adds PDF support for Sonnet and Gemini models, enabling document context in chats.
- ›New
--voice-input-deviceflag lets you select the audio input device for voice recording. - ›New
--timeoutoption configures API call timeouts. - ›Sets working directory to repo root when running shell commands.
- ›Adds Ctrl-Up/Down keyboard shortcuts for per-message history navigation.
+1 moreshow less
- ›Improves
/dropcommand to support substring matching for non-glob patterns.
└──▷ BREAKING ON UPGRADE- !Dart support has been removed.
- v0.65.0
Aider v0.65.0 adds custom model aliases, URL-scraping control, Dart RepoMap support, and an improved
/editorcommand.└──▷ GET THIS VERSION$ git clone --branch v0.65.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.65.0
└──▷ TRY ITPrevent aider from prompting you to scrape URLs that appear in the chat, reducing interruptions during automated or scripted sessions.$ aider --no-detect-urls
- ›Adds
--aliasconfig option to define custom model aliases for frequently used models. - ›Adds
--[no-]detect-urlsflag to enable or disable automatic detection and scraping of URLs found in chat. - ›Adds RepoMap support for the Dart language, enabling code-map navigation in Dart repositories.
- ›Ollama models now default to an 8k context window.
- ›The
/editorcommand now returns and prefills file content into the prompt, enabling composed messages that begin with slash commands.
- ›Adds
- v0.64.0
Aider v0.64.0 adds
/editorcommand, multiline input tags, global extra_params, and GPT-4o/o1 streaming support.└──▷ GET THIS VERSION$ git clone --branch v0.64.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.64.0
└──▷ TRY ITSuppress the release notes pop-up in automated or CI environments where aider is invoked non-interactively.$ aider --show-release-notes=false <file>
Pass model-specific parameters (e.g. temperature) globally to litellm for every completion call without per-model config.# In ~/.aider.model.settings.yml - name: aider/extra_params extra_params: temperature: 0.2 max_tokens: 4096- ›Adds
/editorcommand to open the system editor for composing prompts without leaving the chat. - ›Adds support for optional multiline input tags with matching closing tags for structured multi-line chat messages.
- ›Adds
--show-release-notesflag to control whether release notes are displayed on first run of a new version. - ›Supports global
extra_paramsin model settings configuration, passed directly to litellm.completion(). - ›Architect mode now automatically prompts to add files suggested by the LLM.
+4 moreshow less
- ›Enables streaming for o1 models by default.
- ›Adds full support for
gpt-4o-2024-11-20. - ›Allows recursive directory loading via
--read. - ›Adds exponential backoff retry when writing files to handle editor file locks.
- ›Adds
- v0.63.0
Aider v0.63.0 adds Qwen 2.5 Coder 32B support and a quieter
/webcommand that appends pages without triggering LLM responses.└──▷ GET THIS VERSION$ git clone --branch v0.63.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.63.0
└──▷ TRY ITFetch a reference page (e.g., an API doc) into context without immediately prompting the LLM — useful when you want to stage multiple sources before asking a question.$ /web https://example.com/api-reference- ›Supports Qwen 2.5 Coder 32B as a model option.
- ›The
/webcommand now adds a page to the chat context silently, without triggering an LLM response. - ›Improved prompting for the user's preferred chat language.
- v0.62.0
Aider v0.62.0 adds Claude 3.5 Haiku support and a new
--apply-clipboard-editsflag to apply LLM edits from web apps.└──▷ GET THIS VERSION$ git clone --branch v0.62.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.62.0
└──▷ TRY ITUse Claude 3.5 Haiku as a cost-effective alternative to Sonnet for code editing sessions.$ aider --haiku
Apply LLM-generated edits copied from a web app (e.g. ChatGPT or Claude.ai) directly to a local file without re-entering the prompt in Aider.$ aider --apply-clipboard-edits file-to-edit.js
- ›Adds
--haikuflag to launch Aider using Claude 3.5 Haiku (75% on the code editing leaderboard, lower cost than Sonnet). - ›New
--apply-clipboard-editsflag applies file changes copied from ChatGPT, Claude, or other web app LLM responses directly to local files.
- ›Adds
- v0.61.0
Aider v0.61.0 adds saveable chat context files, opt-in analytics, and new launch/input flags.
└──▷ GET THIS VERSION$ git clone --branch v0.61.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.61.0
└──▷ TRY ITSnapshot your current file context so you can restore it in a future session or share it with a teammate.$ aider> /save context.aider # later or in another session: aider> /load context.aiderPre-load a standard set of project files and slash-commands before the interactive session starts — useful for CI-style automation.$ aider --load project_context.aider --model gpt-4o
Run aider in environments where prompt toolkit causes display issues, disabling fancy input while keeping output formatting.$ aider --no-fancy-input
- ›New
/save <fname>command exports current chat file context as a replayable file of/addand/read-onlycommands. - ›New
/load <fname>command replays any set of slash-commands from a file, enabling scripted chat workflows. - ›New
--load <fname>flag runs a command file at launch, before interactive chat begins. - ›Adds anonymous, opt-in analytics with no personal data sharing.
- ›Enables image support for models dynamically by following litellm's
supports_visionattribute.
+4 moreshow less
- ›New
--no-fancy-inputswitch disables prompt toolkit input while keeping other UI features active. - ›New
--no-browser/--no-guiflags override browser/GUI config at launch. - ›Offers to open documentation URLs automatically when errors occur.
- ›Adds full support for all o1 models regardless of provider.
- ›New
- v0.60.0
Aider v0.60.0 defaults to Claude Sonnet 10/22 and adds bundled model metadata for faster LLM support.
└──▷ GET THIS VERSION$ git clone --branch v0.60.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.60.0
- ›Adds full support for Claude Sonnet 10/22 (new SOTA on aider's code editing benchmark) and sets it as the default model.
- ›Improves o1 model support with more flexible parsing of nonconforming code edit replies.
- ›Bundles a model metadata JSON file to support models not yet updated in litellm.
- ›Adds model settings for o1 models on Azure.
- ›Automatically offers to add
.envto.gitignorewhen detected.
- v0.59.0
Aider v0.59.0 adds glob support for
/read-only, faster large-repo launch, and editor-model sanity checks.└──▷ GET THIS VERSION$ git clone --branch v0.59.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.59.0
└──▷ TRY ITAdd all Python files under a directory as read-only context without adding them one by one.$ /read-only src/**/*.pySpeed up aider startup when working inside a very large repository where repo validation is slow.$ aider --skip-sanity-check-repo --model gpt-4o
- ›Adds shell-style filesystem auto-complete and glob support (e.g.
src/**/*.py) to the/read-onlycommand. - ›New
--skip-sanity-check-repoflag skips repository sanity checks at launch to speed up startup in large repos. - ›The
/settingscommand now includes the full announcement lines printed at launch for a complete config snapshot. - ›Sanity-checks the
--editor-modelon launch, matching the validation already applied to main and weak models. - ›Repo-map generation is now deterministic with improved caching logic.
└──▷ BREAKING ON UPGRADE- !The
--yesflag is renamed to--yes-always; existing YAML files usingyes:and.envfiles usingAIDER_YESmust be updated toyes-always:andAIDER_YES_ALWAYSrespectively. - !Config files now use standard YAML list syntax (
- list entries, one per line); existing config files using the old list format must be updated.
- ›Adds shell-style filesystem auto-complete and glob support (e.g.
- v0.58.0
Aider v0.58.0 adds Architect/Editor dual-model mode, o1 shortcuts, new
/copycommand, and voice format control.└──▷ GET THIS VERSION$ git clone --branch v0.58.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.58.0
└──▷ TRY ITRecord and send voice input as mp3 instead of the default format.$ aider --voice-format mp3
Quickly start a session using o1-mini without specifying a full model string.$ aider --o1-mini
- ›Adds Architect/Editor dual-model coding mode: pair a strong reasoning model (e.g. o1-preview) as Architect with a faster model (e.g. gpt-4o) as Editor.
- ›New
--o1-previewand--o1-minishortcut flags for quick model selection. - ›Adds support for Gemini 002 and Qwen 2.5 models.
- ›New
/copycommand copies the last LLM response to the clipboard. - ›New
--voice-formatswitch to send voice audio as wav, mp3, or webm.
+6 moreshow less
- ›Adds
extra_paramsdict toModelSettingsfor passing arbitrary extra parameters to litellm.completion(). - ›Many confirmation prompts can now be permanently skipped for the session via a '(D)on't ask again' response.
- ›Autocomplete for
/read-onlynow covers the entire filesystem. - ›New settings for customizing completion menu colors.
- ›Supports cursor shapes when using vim mode.
- ›Aider now follows HTTP redirects when scraping URLs.
└──▷ BREAKING ON UPGRADE- !The
/clipboardcommand has been renamed to/paste.
- v0.57.0
Aider v0.57.0 adds o1-preview/o1-mini support, recursive read-only dirs, and new Cohere models.
└──▷ GET THIS VERSION$ git clone --branch v0.57.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.57.0
└──▷ TRY ITUse OpenAI's o1-preview reasoning model for higher-accuracy code edits in a security-sensitive codebase.$ aider --model o1-preview
- ›Supports OpenAI o1-preview and o1-mini models via
--model o1-previewand--model o1-mini. - ›o1-preview with diff edit format now matches SOTA leaderboard result previously only achieved with whole edit format.
- ›Supports new Cohere 08-2024 models.
- ›Enables recursive directory addition with
/read-only.
- ›Supports OpenAI o1-preview and o1-mini models via
- v0.56.0
Aider v0.56.0 adds
/reportcommand,--chat-languageswitch, and prompt caching for Sonnet via OpenRouter.└──▷ GET THIS VERSION$ git clone --branch v0.56.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.56.0
└──▷ TRY ITFile a bug report instantly from the CLI — opens a pre-filled GitHub Issue in your browser.$ /reportForce all of Aider's chat output into a specific spoken language, useful for non-English teams.$ aider --chat-language Spanish
Disable shell-command suggestions entirely so Aider never prompts to run commands in your session.$ aider --no-suggest-shell-commands
- ›New
/reportcommand opens the browser with a pre-populated GitHub Issue for quick bug reporting. - ›New
--chat-languageswitch sets the spoken language for chat responses. - ›Enables prompt caching for Sonnet via OpenRouter.
- ›Enables 8k output tokens for Sonnet via VertexAI and DeepSeek V2.5.
- ›
--[no-]suggest-shell-commandsnow controls both prompting for and offering to execute shell commands.
+1 moreshow less
- ›Checks key imports on launch and surfaces a helpful error message when dependencies are missing.
└──▷ BREAKING ON UPGRADE- !The
--modelsflag is renamed to--list-models; any scripts or aliases using--modelswill break.
- ›New
- v0.55.0
Aider v0.55.0 adds
/settingscommand, crash reporting, devops repo map, and new color/verbose options.└──▷ GET THIS VERSION$ git clone --branch v0.55.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.55.0
└──▷ TRY ITReview all active Aider settings mid-session to confirm flags and config are applied as expected.$ /settings- ›New
/settingscommand displays all currently active settings at a glance. - ›Includes important devops files (e.g., CI configs) in the repo map for better LLM context.
- ›Adds
--tool-warning-colorsetting to customize warning message color. - ›Adds
--verbosedebug output for shell commands. - ›Shares active test and lint commands with the LLM during sessions.
+4 moreshow less
- ›Adds
--updateas an alias for--upgrade. - ›Refuses to create a git repo in
$HOMEand warns the user instead. - ›Catches
/voicetranscription exceptions and surfaces the WAV file for manual recovery. - ›Skips fuzzy filename matching when the LLM is creating a new file, preventing accidental overwrites.
- ›New
- v0.54.0
Aider v0.54.0 adds interactive shell commands, LLM output sharing, and a self-upgrade switch.
└──▷ GET THIS VERSION$ git clone --branch v0.54.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.54.0
└──▷ TRY ITKeep LLM context fresh by automatically feeding shell command output back to the model during a coding session.$ aider --suggest-shell-commands
Update Aider to the latest PyPI release without leaving the tool or running pip manually.$ aider --upgrade
- ›Shell and
/runcommands are now interactive in pty-capable environments, enabling real-time input during command execution. - ›New
--suggest-shell-commands/--no-suggest-shell-commandsflag controls whether Aider proposes shell commands and optionally feeds their output back to the LLM. - ›New
--upgradeswitch installs the latest Aider release directly from PyPI without leaving the tool. - ›Adds model settings support for
gemini/gemini-1.5-pro-exp-0827andgemini/gemini-1.5-flash-exp-0827. - ›Performance improvements to autocomplete in large and monorepo codebases.
- ›Shell and
- v0.53.0
Aider v0.53.0 adds prompt cache keepalive pings, bulk confirmation controls, and a dev-branch installer.
└──▷ GET THIS VERSION$ git clone --branch v0.53.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.53.0
└──▷ TRY ITPull the latest unreleased Aider changes from the main branch to test cutting-edge features before a formal release.$ aider --install-main-branch
- ›Adds
--cache-keepalive-pingsto periodically ping the API every 5 minutes and prevent prompt cache expiration. - ›Enables bulk accept/reject of add-URL and run-shell confirmation prompts in a single action.
- ›Adds
--install-main-branchflag to upgrade directly to the latest development version of Aider. - ›Surfaces model metadata information when running with
--verbose. - ›Extends
extra_headerssupport to the weak model, enabling Anthropic beta features for that model tier.
- ›Adds
- v0.52.0
Aider v0.52.0 adds interactive shell command suggestions, a new
/resetcommand, and repo map size control.└──▷ GET THIS VERSION$ git clone --branch v0.52.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.52.0
└──▷ TRY ITReset an entire session — drop all tracked files and clear history — to start a clean context mid-conversation.$ /resetWiden the repo map when working without explicitly added files, so aider has more codebase context to draw on.$ aider --map-multiplier-no-files 4
- ›Adds automatic shell command suggestions after edits — offers to launch a browser, install dependencies, run DB migrations, execute the program, or run new tests.
- ›New
/resetcommand drops all files and clears chat history in one step. - ›New
--map-multiplier-no-filesflag controls the repo map size multiplier when no files are loaded in chat. - ›Switches default OpenAI model to
gpt-4o-2024-08-06. - ›
/readand/dropnow expand~to the home directory.
+1 moreshow less
- ›Displays the active chat mode in the aider prompt.
└──▷ BREAKING ON UPGRADE- !The default repo map size multiplier when no files are in chat is reduced to 2 (previously higher); this changes aider's context behavior on upgrade without any config change.
- v0.51.0
Aider v0.51.0 adds Anthropic prompt caching, configurable repo-map refresh, and improved Jupyter Notebook editing.
└──▷ GET THIS VERSION$ git clone --branch v0.51.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.51.0
└──▷ TRY ITPrevent repo map recomputation on every interaction in a monorepo — recompute only when tracked files change.$ aider --map-refresh files
- ›Adds
--cache-promptsflag to enable prompt caching for Anthropic models, covering the system prompt, repo map, and/read-onlyfiles. - ›Adds
--map-refreshflag with modesalways,files,manual, andautoto control how often the repo map recomputes — useful for large or monorepos. - ›Improves editing performance on Jupyter Notebook
.ipynbfiles. - ›Shows which config YAML file is loaded when running with
--verbose.
- ›Adds
- v0.50.0
Aider v0.50.0 adds DeepSeek Coder support, infinite output for more models, and new chat-mode switching.
└──▷ GET THIS VERSION$ git clone --branch v0.50.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.50.0
└──▷ TRY ITUse DeepSeek Coder as your backend for a coding session with extended 8k-token output.$ aider --deepseek
Launch directly into ask mode for exploratory Q&A, then request a code edit in-place without switching tools.$ aider --chat-mode ask # then inside the session: /code refactor the authentication module to use bcrypt
- ›New
--deepseekswitch enables DeepSeek Coder with 8k token output. - ›Infinite output (streaming past normal limits) now supported for DeepSeek Coder and Mistral models, in addition to Anthropic models.
- ›New
--chat-mode <mode>switch lets you launch directly intoask,help, orcodemode. - ›New
/code <message>command lets you request a code edit without leavingaskmode.
- ›New
- v0.49.0
Aider v0.49.0 adds read-only file context, clipboard pasting, in-script slash commands, and new model shortcuts.
└──▷ GET THIS VERSION$ git clone --branch v0.49.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.49.0
└──▷ TRY ITInclude a read-only reference file (e.g., a shared schema outside your repo) so aider can see it without editing it.$ aider --read ../shared/schema.sql src/models.py
Paste a screenshot or copied code from your clipboard directly into an aider session for analysis or editing.$ /clipboardRun aider non-interactively in a CI script and include slash commands in the message to control chat behavior.$ aider --message "/read ../policy.md Summarize any security policy violations in auth.py" auth.py
- ›Adds
/readcommand and--readflag to include read-only files in chat context, including files outside the git repo. - ›New
/clipboardcommand pastes images or text from the clipboard directly into the chat (replaces/add-clipboard-image). - ›Enables in-chat
/commands inside scripted aider messages for automation workflows. - ›Adds
--miniflag as a shorthand to usegpt-4o-mini. - ›Adds
--attribute-commit-message-authorto prefix commit messages with 'aider: ' only when aider authored the changes.
+5 moreshow less
- ›Adds
--attribute-commit-message-committerto prefix all commit messages with 'aider: ' regardless of author. - ›Adds support for
openai/gpt-4o-2024-08-06model. - ›
/tokensnow displays the active model alongside token counts. - ›Shows scraped markdown content when a URL is added via
/web. - ›Removes the obsolete 16k token hard limit on commit diffs, deferring to per-model limits instead.
└──▷ BREAKING ON UPGRADE- !
--attribute-commit-messageis replaced by--attribute-commit-message-author; setups using--attribute-commit-messagemust migrate to the new flag.
- ›Adds
- v0.48.0
Aider v0.48.0 adds clipboard image input, subtree scoping, and unlocks repo maps for any model.
└──▷ GET THIS VERSION$ git clone --branch v0.48.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.48.0
└──▷ TRY ITScope aider to a subdirectory of a large monorepo to avoid indexing the entire repo and speed up context building.$ cd services/auth && aider --subtree-only
Enable repo map on a model that wouldn't use one by default, giving it structural context of your codebase.$ aider --map-tokens 1024 --model <your-model>
Add a screenshot or diagram from your clipboard into the chat to give aider visual context.$ /add-clipboard-image- ›New
/add-clipboard-imagecommand lets you paste images directly into the chat. - ›New
--subtree-onlyflag scopes aider to the current directory subtree, improving large and mono-repo performance. - ›Use
--map-tokens 1024to enable repo map with any model, not just defaults. - ›Supports Sonnet's 8k output window.
- ›New
- v0.47.0
Aider v0.47.0 adds customizable commit prompts, Conventional Commits support, a full Docker image, and generic command auto-completions.
└──▷ GET THIS VERSION$ git clone --branch v0.47.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.47.0
└──▷ TRY ITEnforce a project-specific commit style (e.g., Jira ticket prefix) instead of the default prompt.$ aider --commit-prompt 'Write a commit message prefixed with the Jira ticket number from the branch name, followed by a short imperative summary.'
Run aider as a non-root user inside Docker without permission errors, using the full-featured image.$ docker run --user $(id -u):$(id -g) -v $(pwd):/app paulgauthier/aider-full
- ›Adds
--commit-promptflag to fully customize the commit message prompt. - ›Adds Conventional Commits guidelines to the default commit message prompt.
- ›Uses the strong model as a fallback for generating commit messages and chat summaries.
- ›Introduces
paulgauthier/aider-fullDocker image bundling all extras. - ›Provides generic auto-completions for
/commandsthat lack a dedicated completion override.
+1 moreshow less
- ›Switching to code or ask mode no longer triggers a chat history summarization.
- ›Adds
- v0.46.0
Aider v0.46.0 adds
/askand/chat-modecommands, always-load file config, and improved token cost reporting.└──▷ GET THIS VERSION$ git clone --branch v0.46.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.46.0
└──▷ TRY ITAsk a question about your code without triggering any edits — useful for reconnaissance before making changes.$ aider /ask What does the authentication flow in auth.py do?Always load a conventions or style-guide file into every session so the model follows project standards automatically.# .aider.conf.yml file: CONVENTIONS.md
Switch to help mode mid-session to troubleshoot aider configuration without leaving the chat.$ /chat-mode help- ›New
/ask <question>command lets you query your codebase without triggering any edits. - ›New
/chat-mode <mode>command switches betweenask,code, andhelpinteraction modes mid-session. - ›Supports
file:key in.aider.conf.ymlto always load one or more files into every chat session. - ›Enhanced token usage and cost reporting now works during streaming responses.
- ›Filename auto-complete for
/addand/dropis now case-insensitive.
+1 moreshow less
- ›
--no-verify-sslnow also disables SSL certificate verification when/webscrapes websites.
- ›New
- v0.45.0
Aider v0.45.0 adds GPT-4o mini support and exposes last 4 chars of API keys in verbose output.
└──▷ GET THIS VERSION$ git clone --branch v0.45.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.45.0
- ›Supports GPT-4o mini, which scores comparably to GPT-3.5 on the code editing benchmark using whole edit format.
- ›Improved file-add suggestions on Windows when aider detects relevant files not yet in the chat.
- ›Shows last 4 characters of API keys in
--verboseoutput for easier key identification.
- v0.44.0
Aider v0.44.0 slashes install size by up to 12x and introduces optional package extras for help, browser, and Playwright.
└──▷ GET THIS VERSION$ git clone --branch v0.44.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.44.0
└──▷ TRY ITCheck for a new Aider release without starting an interactive session, using the renamed flag.$ aider --just-check-updated
- ›Reduces default pip install size by 3–12x, making minimal installs practical in constrained environments.
- ›Adds three optional package extras (
aider-chat[help],aider-chat[browser],aider-chat[playwright]) that Aider will prompt to install on demand. - ›Simplifies the output of
--modelsfor easier model browsing.
└──▷ BREAKING ON UPGRADE- !The
--check-updateflag is renamed to--just-check-updated; any scripts or aliases using--check-updatewill break. - !The
--skip-check-updateflag is renamed to--[no-]check-update; any scripts or aliases using--skip-check-updatewill break.
- v0.43.0
Aider v0.43.0 adds in-chat help, multi-undo, and standardized config file loading from multiple paths.
└──▷ GET THIS VERSION$ git clone --branch v0.43.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.43.0
└──▷ TRY ITGet contextual help on configuring a setting or troubleshooting an issue without leaving the chat session.$ /help How do I set a custom model settings file?- ›Adds
/help <question>command to ask questions about usage, settings, troubleshooting, and LLM configuration directly inside the chat session. - ›Default
--model-settings-filepath is now.aider.model.settings.yml. - ›Default
--model-metadata-filepath is now.aider.model.metadata.json. - ›All config, env, YAML, and JSON files now load from home directory, git root, current working directory, and a named command-line switch — enabling layered configuration.
- ›Introduces
$HOME/.aider/cachesas a new directory for app-wide expendable caches.
+1 moreshow less
- ›Allows multiple sequential uses of
/undoto step back through multiple changes.
└──▷ BREAKING ON UPGRADE- !The default value of
--model-settings-fileis now.aider.model.settings.yml; setups relying on a previous default path will no longer load automatically. - !The default value of
--model-metadata-fileis now.aider.model.metadata.json; setups relying on a previous default path will no longer load automatically.
- ›Adds
- v0.41.0
Aider v0.41.0 defaults to Claude 3.5 Sonnet, adds >4k token streaming, image support, and commit message prefixing.
└──▷ GET THIS VERSION$ git clone --branch v0.41.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.41.0
└──▷ TRY ITPrefix all aider-generated commits with 'aider:' to make AI-assisted changes easy to identify in git history.$ aider --attribute-commit-message
- ›Adds
--attribute-commit-messageflag to prefix aider's commit messages with 'aider:'. - ›Aider now defaults to
claude-3-5-sonnet-20240620whenANTHROPIC_API_KEYis set in the environment. - ›Enables streaming responses greater than 4k tokens for Claude 3.5 Sonnet, unlocking large refactors and multi-file code generation in a single pass.
- ›Adds image support for Claude 3.5 Sonnet and for GPT-4o and Claude 3.5 Sonnet via OpenRouter.
- ›Automatically retries requests on Anthropic
overloaded_errorresponses.
- ›Adds
- v0.40.0
Aider v0.40.0 adds per-model settings file and options to suppress git commit attribution tagging.
└──▷ GET THIS VERSION$ git clone --branch v0.40.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.40.0
└──▷ TRY ITApply custom repo-map and token settings for a specific model without modifying global config.$ aider --model-settings-file ./my-model-settings.yml
- ›Adds
--model-settings-fileflag to supply a custom per-model settings file, enabling per-model control over behaviours such as repo-map usage. - ›Adds options to suppress appending '(aider)' to git author and committer names in commit attribution.
- ›Adds
- v0.39.0
Aider v0.39.0 adds Claude 3.5 Sonnet support, LLM message logging, and .env-based config for all AIDER_ variables.
└──▷ GET THIS VERSION$ git clone --branch v0.39.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.39.0
└──▷ TRY ITRun Aider with the top-ranked Claude 3.5 Sonnet model for highest-quality code edits.$ aider --sonnet
Capture every raw message sent to the LLM to a file for auditing or debugging prompt behaviour.$ aider --sonnet --llm-history-file llm-messages.log
- ›Adds
--sonnetflag to select Claude 3.5 Sonnet as the active LLM model. - ›Adds
--llm-history-fileflag to log raw messages sent to the LLM for inspection or auditing. - ›All
AIDER_xxxenvironment variables can now be set in a.envfile instead of the shell environment. - ›Git commit author and committer names now include '(aider)' instead of prefixing commit messages with 'aider:'.
└──▷ BREAKING ON UPGRADE- !Commit messages are no longer prefixed with 'aider:' — the git author and committer names now have '(aider)' appended instead. Any tooling or scripts that filter or parse aider-generated commits by the 'aider:' message prefix will need to be updated.
- ›Adds
- v0.38.0
Aider v0.38.0 adds
--vimkeybindings and per-model metadata via.aider.models.json.└──▷ GET THIS VERSION$ git clone --branch v0.38.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.38.0
└──▷ TRY ITUse vim keybindings when working in the Aider chat for faster navigation and editing.$ aider --vim
Override context window size and token costs for a custom or private LLM endpoint that Aider does not know about natively.{ "my-custom-model": { "max_tokens": 8192, "input_cost_per_token": 0.000002, "output_cost_per_token": 0.000006 } }- ›Adds
--vimflag to enable vim/vi keybindings in the interactive chat session. - ›Adds
.aider.models.jsonfile support to specify custom LLM metadata such as context window size and token costs. - ›Ensures
--commit --dry-runis a true no-op, making dry-run safe to use in automation. - ›Delivers more detailed error messages when token limit errors occur.
- ›Adds
- v0.37.0
Aider v0.37.0 adds URL scraping from chat, smarter repo map optimization, and audio sample rate detection for
/voice.└──▷ GET THIS VERSION$ git clone --branch v0.37.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.37.0
- ›Repo map is now optimized based on the text of chat history in addition to files added to chat, improving context relevance.
- ›Aider detects and offers to scrape URLs pasted directly into the chat.
- ›Detects supported audio sample rates for the
/voicecommand. - ›Improved prompts when no files are added to chat, encouraging the LLM to suggest relevant files.
- ›Performance improvements to repo map generation, especially in large repos.
+1 moreshow less
- ›Overrides
GIT_EDITORenvironment variable for commits even when it is already set in the environment.
- v0.36.0
Aider v0.36.0 adds automatic lint-and-fix after every LLM edit plus integrated test-run-and-fix support.
└──▷ GET THIS VERSION$ git clone --branch v0.36.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.36.0
└──▷ TRY ITRun aider with a custom linter (e.g. flake8) so every LLM edit is automatically linted and fixed before aider finishes its turn.$ aider --lint-cmd 'flake8 --max-line-length 120' mymodule.py
Wire up your test suite so aider auto-fixes failures after each change — useful for TDD workflows where you want the AI to keep iterating until tests pass.$ aider --test-cmd 'pytest tests/' --test mymodule.py
- ›Adds
--lint-cmdflag to configure a custom linter, with built-in basic linting for all tree-sitter-supported languages when no external linter is set. - ›Adds
/lintchat command and--lintCLI flag to manually trigger lint-and-fix on files; aider also runs this automatically after every LLM edit. - ›Adds
--test-cmdflag to configure a test runner command,/testchat command and--testCLI flag to run tests on demand; aider automatically attempts to fix any test failures.
- ›Adds
- v0.35.0
Aider v0.35.0 switches to GPT-4o as the default model and adds
--restore-chat-historyto resume prior sessions.└──▷ GET THIS VERSION$ git clone --branch v0.35.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.35.0
└──▷ TRY ITResume a previous coding session without re-explaining context to the model.$ aider --restore-chat-history
- ›Adds
--restore-chat-historyflag to restore prior chat history on launch, enabling continuation of the last conversation. - ›Switches the default model to GPT-4o, which scores 72.9% on the aider LLM code editing leaderboard (up from 68.4% for Opus).
- ›Improves reflection feedback to LLMs when using the diff edit format.
- ›Adds
- v0.34.0
Aider v0.34.0 adds
--show-promptsdebug switch and explicit model settings for Claude 3 Opus and GPT-3.5 Turbo.└──▷ GET THIS VERSION$ git clone --branch v0.34.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.34.0
└──▷ TRY ITInspect the exact prompts Aider sends to the model to debug unexpected edits or tune behavior.$ aider --show-prompts
- ›Adds
--show-promptsdebug switch to inspect the prompts sent to the model. - ›Adds explicit model settings for
openrouter/anthropic/claude-3-opusandgpt-3.5-turbo.
- ›Adds
- v0.32.0
Aider v0.32.0 adds LLM code-editing leaderboards, a new diff-fenced edit format for Gemini 1.5 Pro, and Deepseek-V2 support.
└──▷ GET THIS VERSION$ git clone --branch v0.32.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.32.0
- ›Gemini 1.5 Pro now defaults to the
diff-fencededit format, improving performance on larger codebases. - ›Adds support for Deepseek-V2 via more flexible system-message configuration in the diff edit format.
- ›Benchmark output now serializes results in YAML format, compatible with the new leaderboard.
- ›Publishes LLM code-editing leaderboards ranking GPT-3.5/4 Turbo, Opus, Sonnet, Gemini 1.5 Pro, Llama 3, Deepseek Coder, and Command-R+ by code-editing ability.
- ›Gemini 1.5 Pro now defaults to the
- v0.31.0
Aider v0.31.0 adds a browser UI mode and in-chat model switching.
└──▷ GET THIS VERSION$ git clone --branch v0.31.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.31.0
└──▷ TRY ITStart Aider in browser mode for a point-and-click pair-programming session instead of the terminal.$ aider --browser
- ›Adds
--browserflag to launch an experimental browser-based UI instead of the terminal interface. - ›Adds
/model <name>chat command to switch the active AI model mid-session. - ›Adds
/models <query>chat command to search the list of available models.
- ›Adds
- v0.30.0
Aider v0.30.0 adds Gemini 1.5 Pro and Groq Llama3 70B support, plus new model search and warning flags.
└──▷ GET THIS VERSION$ git clone --branch v0.30.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.30.0
└──▷ TRY ITDiscover available model names matching a pattern before starting a session.$ aider --models gemini
Suppress model-compatibility warnings when using a lesser-known or custom model in CI pipelines.$ aider --model groq/llama3-70b-8192 --no-show-model-warnings
- ›Adds
--models <MODEL-NAME>flag to search available models by name. - ›Adds
--no-show-model-warningsflag to silence warnings about unknown or unfamiliar models. - ›Enables repo map for the 'whole' edit format, expanding context awareness to that mode.
- ›Adds Gemini 1.5 Pro as a recommended free model.
- ›Adds improved support for Groq's Llama3 70B model.
- ›Adds
- v0.29.0
Aider v0.29.0 adds direct LLM provider connections and new model-selection flags including
--opus,--sonnet, and--weak-model.└──▷ GET THIS VERSION$ git clone --branch v0.29.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.29.0
└──▷ TRY ITUse a cheaper model just for commit messages and history summarization while keeping a powerful model for edits.$ aider --opus --weak-model claude-3-haiku-20240307
Connect to Anthropic's Sonnet model directly by supplying your API key on the CLI.$ aider --sonnet --anthropic-api-key <your-api-key>
- ›Adds
--weak-model <model-name>flag to specify which model handles commit messages and chat history summarization independently of the main model. - ›Adds
--opusand--sonnetCLI flags for direct Anthropic model selection. - ›Adds
--4-turbo-visionCLI flag for GPT-4 Turbo with vision support. - ›Adds
--anthropic-api-keyCLI flag to supply an Anthropic API key directly. - ›Adds direct connection support for Anthropic, Cohere, Gemini, and many other LLM providers.
+2 moreshow less
- ›Improves 'whole' and 'diff' backends to better support Cohere's Command-R+ model.
- ›Allows
/addof images from anywhere in the filesystem, not just within the repo.
- ›Adds
- v0.28.0
Aider v0.28.0 adds support for
gpt-4-turbo-2024-04-09andgpt-4-turbomodels.└──▷ GET THIS VERSION$ git clone --branch v0.28.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.28.0
└──▷ TRY ITSwitch to the newly supported gpt-4-turbo model for a coding session.$ aider --model gpt-4-turbo
- ›Adds support for the
gpt-4-turbo-2024-04-09andgpt-4-turbomodels (61.7% on Exercism benchmark, 34.1% on refactoring/laziness benchmark); default remainsgpt-4-1106-preview.
- ›Adds support for the
- v0.26.0
Aider v0.26.0 defaults to GPT-4 Turbo and adds
-3/-4model-selection switches.└──▷ GET THIS VERSION$ git clone --branch v0.26.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.26.0
└──▷ TRY ITForce GPT-3.5 for a faster, cheaper session when full GPT-4 power isn't needed.$ aider -3 <file>
Pin to classic GPT-4 (non-Turbo) when you need its specific behaviour rather than the new default Turbo model.$ aider -4 <file>
- ›Adds
-4switch to select GPT-4 (non-Turbo) and-3switch to select GPT-3.5 at launch. - ›Switches the default model to GPT-4 Turbo.
- ›Adds
- v0.25.0
Aider v0.25.0 adds
--openai-organization-id, chat size warnings, and.aiderignoreenforcement└──▷ GET THIS VERSION$ git clone --branch v0.25.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.25.0
└──▷ TRY ITAuthenticate Aider against a specific OpenAI organization account, useful when your API key belongs to multiple orgs.$ aider --openai-organization-id org-xxxxxxxxxxxxxxxx
- ›Adds
--openai-organization-idCLI argument to specify an OpenAI organization when invoking Aider. - ›Enforces
.aiderignoreat chat-add time — vocally refuses to add matching files, preventing downstream git commit failures. - ›Issues a warning when too much code is added to the chat, with a link to the FAQ on managing chat context.
- ›Surfaces a FAQ link when edits fail to apply, guiding users toward resolution steps.
- ›Adds
- v0.24.0
Aider v0.24.0 adds
/webURL scraping and updates default GPT-3.5 model to gpt-3.5-turbo-0125.└──▷ GET THIS VERSION$ git clone --branch v0.24.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.24.0
└──▷ TRY ITPull a CVE advisory or documentation page directly into an Aider session so the model can reason about its content.$ /web https://nvd.nist.gov/vuln/detail/CVE-2024-12345- ›Adds
/web <url>command that scrapes a URL, converts the page to clean markdown, and injects it into the chat context. - ›Changes the default GPT-3.5 model to
gpt-3.5-turbo-0125. - ›Updates all OpenAI model names and pricing information.
- ›Adds
- v0.23.0
Aider v0.23.0 adds
/testcommand,--openroutershortcut, and new model aliases for GPT-4 Turbo.└──▷ GET THIS VERSION$ git clone --branch v0.23.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.23.0
└──▷ TRY ITRun your test suite from within a chat session and automatically surface failures to the model for fixing.$ /test pytest tests/Route all requests through OpenRouter without manually specifying the full base URL.$ aider --openrouter --model mistralai/mistral-7b-instruct
- ›Adds
--model gpt-4-0125-previewand--model gpt-4-turbo-preview(OpenAI alias) as supported model targets;--4turboremains an alias for--model gpt-4-1106-preview. - ›Adds
--openrouterflag as a shortcut for--openai-api-base https://openrouter.ai/api/v1, enabling one-flag OpenRouter integration. - ›Adds
--skip-check-updateflag to suppress the update check on launch. - ›New
/testcommand runs an arbitrary command and automatically appends its output to the chat when the exit status is non-zero. - ›Adds
/quitas an alias for/exit.
- ›Adds
- v0.22.0
Aider v0.22.0 adds
!as a shell-run alias and improves unified diff editing and autocomplete.└──▷ GET THIS VERSION$ git clone --branch v0.22.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.22.0
└──▷ TRY ITRun a shell command directly from the Aider prompt without typing/run.$ ! pytest tests/- ›Adds
!as an alias for the/runcommand, providing a shorthand to execute shell commands from the chat prompt. - ›Autocomplete for
/addand/dropnow properly quotes filenames that contain spaces. - ›The
/undocommand now instructs GPT not to simply retry the reverted edit. - ›Improvements to the unified diff editing format.
- ›Adds
- v0.20.0
Aider v0.20.0 adds image support in chat to automatically invoke GPT-4 Vision.
└──▷ GET THIS VERSION$ git clone --branch v0.20.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.20.0
- ›Adds image attachment support in the chat interface, automatically switching to GPT-4 Vision when an image is included.
- v0.19.0
Aider v0.19.0 adds
--4-turboshortcut and switches GPT-4 Turbo to unified diff format by default└──▷ GET THIS VERSION$ git clone --branch v0.19.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.19.0
└──▷ TRY ITQuickly start an Aider session using GPT-4 Turbo without typing the full model name.$ aider --4-turbo
- ›Adds
--4-turboCLI flag as a shortcut for--model gpt-4-1106-preview. - ›Switches
gpt-4-1106-previewto use unified diffs as the default edit format, reducing incomplete ('lazy') code output — laziness benchmark improves from 20% to 61%.
- ›Adds
- v0.17.0
Aider v0.17.0 adds gpt-4-1106-preview (128k) and gpt-3.5-turbo-1106 (16k) support plus a streamlined scripting API.
└──▷ GET THIS VERSION$ git clone --branch v0.17.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.17.0
- ›Adds support for
gpt-4-1106-previewwith a 128k context window. - ›Adds support for
gpt-3.5-turbo-1106with a 16k context window. - ›Streamlined API for scripting Aider programmatically, with new documentation.
- ›Improved repo-map support for Elisp files.
- ›Adds support for
- v0.16.0
Aider v0.16.0 upgrades repo mapping with tree-sitter and switches to search/replace edit blocks.
└──▷ GET THIS VERSION$ git clone --branch v0.16.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.16.0
- ›Switches the edit format from 'edit block' to 'search/replace block', reducing malformed edits (benchmarked at 66.2%).
- ›Improves repository map generation using tree-sitter for more accurate code context.
- ›Improves handling of malformed edit blocks that target multiple edits to the same file (benchmarked at 65.4%).
- v0.15.0
Aider v0.15.0 adds
.aiderignoresupport, a--commitflag, and full shell commands in/runand/git.└──▷ GET THIS VERSION$ git clone --branch v0.15.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.15.0
└──▷ TRY ITAutomatically commit all staged changes with an AI-generated message as part of a CI step.$ aider --commit
Run a multi-step shell expression inside an aider session to inspect a subdirectory without leaving the chat.$ /run (cd subdir; ls)- ›Adds
.aiderignorefile support to instruct aider to ignore parts of the git repo during context gathering. - ›
/runand/gitslash commands now accept full shell expressions, including subshells like/run(cd subdir; ls). - ›Adds universal ctags and multi-architecture support to the aider Docker image.
- ›Adds
- v0.14.2
Aider v0.14.2 adds Docker support, OpenRouter integration for Claude 2, and improved
/addand/dropfile handling.└──▷ GET THIS VERSION$ git clone --branch v0.14.2 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.14.2
- ›Supports running Aider directly from a Docker image via the published Docker workflow.
- ›Adds Claude 2 and other LLM support through OpenRouter integration.
- ›Extends
/addand/dropcommands to handle absolute filenames and quoted filenames. - ›Adds a guard in
/addto verify files are within the git repo or root before adding them. - ›Shows models supported by the API/key when the requested model is unavailable.
+1 moreshow less
- ›Warns users when in-chat file paths must be relative to the git repo root.
└──▷ BREAKING ON UPGRADE- !Aider now requires Python >= 3.9; installations running Python 3.8 or earlier will no longer work.
- v0.14.1
Aider v0.14.1 improves
/addand/droppath handling and shows available models when a requested model is unavailable.└──▷ GET THIS VERSION$ git clone --branch v0.14.1 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.14.1
- ›The
/addand/dropcommands now accept absolute filenames and quoted filenames. - ›
/addvalidates that files are within the git repo or root before adding them. - ›Shows models supported by the API/key when the requested model is not available.
- ›Warns users when in-chat file paths are relative to the git repo root.
- ›The
- v0.14.0
Aider v0.14.0 adds Claude 2 and multi-LLM support via OpenRouter and requires Python 3.9+.
└──▷ GET THIS VERSION$ git clone --branch v0.14.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.14.0
- ›Adds support for Claude 2 and other LLMs through OpenRouter integration, expanding beyond OpenAI models.
- ›Publishes documentation for running the aider benchmarking suite.
└──▷ BREAKING ON UPGRADE- !Aider now requires Python >= 3.9; installations running Python 3.8 or earlier will no longer work.
- v0.13.0
Aider v0.13.0 adds
--voice-languagefor targeted voice transcription and smarter git commit scoping.└──▷ GET THIS VERSION$ git clone --branch v0.13.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.13.0
└──▷ TRY ITConstrain voice transcription to a specific language to avoid Whisper guessing wrong when working in a non-English codebase.$ aider --voice-language fr
- ›Adds
--voice-languageswitch to constrain the/voicecommand to transcribe to a specific language via Whisper. - ›Sends chat history as prompt/context for Whisper voice transcription, improving accuracy.
- ›Git commits now only include dirty files that GPT actually tries to edit, reducing unintended staging.
- ›Adds
- v0.12.0
Aider v0.12.0 adds voice-to-code support so you can dictate code changes aloud.
└──▷ GET THIS VERSION$ git clone --branch v0.12.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.12.0
- ›Adds voice-to-code support, enabling practitioners to drive code edits through spoken input.
- v0.11.1
Aider v0.11.1 adds a repo-map progress bar and auto-summarizes chat history to prevent context window exhaustion.
└──▷ GET THIS VERSION$ git clone --branch v0.11.1 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.11.1
- ›Adds a progress bar when initially creating a repo map.
- ›Shows more detail on dollar costs when running with
--no-stream.
- v0.11.0
Aider v0.11.0 adds automatic chat history summarization and richer cost reporting with
--no-stream.└──▷ GET THIS VERSION$ git clone --branch v0.11.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.11.0
└──▷ TRY ITSee detailed token cost breakdowns per exchange when running long sessions without streaming.$ aider --no-stream
- ›Adds
--no-streammode with more detailed dollar cost breakdowns per interaction.
- ›Adds
- v0.10.1
Aider v0.10.1 adds
/gitcommand, multiline input via Meta-ENTER, and auto-.gitignorefor aider files.└──▷ GET THIS VERSION$ git clone --branch v0.10.1 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.10.1
└──▷ TRY ITRun a git log or status check without leaving your aider chat session.$ /git log --oneline -5
- ›Adds
/gitcommand to run git commands directly from inside an aider chat session. - ›Adds
/addand/dropcommands now always resolve paths relative to the git root for consistency. - ›Supports multiline chat messages via Meta-ENTER (or Esc+ENTER in some environments).
- ›Adds version-check against PyPI on startup, notifying users when a newer release is available.
- ›Forces
--no-prettymode automatically when aider detects it is running inside a VSCode terminal.
- ›Adds
- v0.10.0
Aider v0.10.0 adds
/gitin-chat command, multiline input via Meta-ENTER, and auto.gitignorefor aider files.└──▷ GET THIS VERSION$ git clone --branch v0.10.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.10.0
└──▷ TRY ITRun a git log or status check without leaving your aider session.$ /git log --oneline -10
- ›Adds
/gitcommand to run git commands directly from inside an aider chat session. - ›Adds
--no-prettyauto-detection: forces--no-prettymode when aider detects it is running inside a VSCode terminal. - ›Supports multiline chat messages via Meta-ENTER (or Esc+ENTER in some environments).
- ›Checks PyPI for newer versions on startup and notifies the user when an update is available.
- ›Updates keyboard interrupt logic so two Ctrl-C presses within two seconds always force-exits aider.
- ›Adds
- v0.9.0
Aider v0.9.0 adds Azure OpenAI support and a new
--show-repo-mapflag.└──▷ GET THIS VERSION$ git clone --branch v0.9.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.9.0
└──▷ TRY ITInspect the repo map Aider will send to the model before starting a session.$ aider --show-repo-map
- ›Adds
--show-repo-mapflag to display the repository map. - ›Supports OpenAI models hosted on Azure.
- ›Adds
- v0.8.3
Aider v0.8.3 adds
--dark-mode,--light-mode, and--encodingflags plus glob support in/addand/drop.└──▷ GET THIS VERSION$ git clone --branch v0.8.3 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.8.3
└──▷ TRY ITLaunch Aider with a dark terminal background so syntax colors remain legible.$ aider --dark-mode
Add all Python files in a directory at once using glob wildcards instead of listing files individually.$ /add src/**/*.pyWork with a codebase that uses a non-UTF-8 encoding, such as Latin-1 source files.$ aider --encoding latin-1
- ›Adds
--dark-modeand--light-modeCLI flags to select terminal colors optimized for dark or light backgrounds. - ›Adds
--encodingflag to specify file encoding (defaults to utf-8); also passed into ctags which is required to return utf-8. - ›Adds
--code-themeflag to control pygments syntax highlighting style of code blocks. - ›Enables glob wildcard support in the
/addand/dropcommands for flexible file selection. - ›Prompts to create a git repository when none is found, enabling GPT code-change tracking from the start.
- ›Adds
- v0.8.2
Aider v0.8.2 adds glob support in
/addand/drop, new--encodingand--code-themeflags, and automatic git repo prompting.└──▷ GET THIS VERSION$ git clone --branch v0.8.2 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.8.2
└──▷ TRY ITScan a project with a non-utf-8 codebase by specifying the correct encoding at startup.$ aider --encoding latin-1 src/legacy_module.py
Add all Python files in a directory at once using glob wildcards instead of listing them individually.$ /add src/**/*.pyApply a dark pygments theme to code blocks for easier reading in light terminals.$ aider --code-theme monokai main.py
- ›Adds
--encodingflag to specify file encoding (defaults to utf-8); also passed through to ctags, which must return utf-8. - ›Adds
--code-themeflag to control pygments styling of code blocks. - ›Supports glob wildcards in
/addand/dropcommands for bulk file management. - ›Prompts to create a git repo when none is found, enabling GPT code-change tracking from first use.
- ›Adds
- v0.8.1
Aider v0.8.1 adds git repo auto-init, glob support in
/addand/drop, and an--encodingflag for ctags.└──▷ GET THIS VERSION$ git clone --branch v0.8.1 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.8.1
└──▷ TRY ITAdd all Python files in a subdirectory at once using glob wildcards instead of listing them individually.$ /add src/**/*.pyWork with a codebase that uses a non-UTF-8 encoding, ensuring files are read and written correctly.$ aider --encoding latin-1 myfile.c
- ›Supports glob wildcards in
/addand/dropcommands, enabling bulk file management in a session. - ›Passes
--encodinginto ctags and requiresutf-8output, improving reliability of code indexing. - ›Prompts to create a git repo when none is found, so GPT code changes are tracked from the start.
- ›Adds
--code-themeswitch to control pygments styling of code blocks. - ›Adds
--encodingswitch to specify a non-default text encoding when reading/writing files.
- ›Supports glob wildcards in
- v0.8.0
Aider v0.8.0 adds
--encodingand--code-themeswitches plus improved Windows support.└──▷ GET THIS VERSION$ git clone --branch v0.8.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.8.0
└──▷ TRY ITScan a codebase on a Windows machine that uses a non-UTF-8 locale, specifying the correct encoding to avoid decoding errors.$ aider --encoding cp1252 myfile.py
Customize the syntax-highlighting theme for code blocks to match a dark terminal palette.$ aider --code-theme monokai myfile.py
- ›Adds
--encodingswitch to specify the character encoding used when reading and writing text files (defaults to utf-8). - ›Adds
--code-themeswitch to control pygments syntax-highlighting style for code blocks in the terminal. - ›Reads and writes text files explicitly with utf-8 encoding by default, benefiting Windows users.
- ›Improves Windows support with path-separator bug fixes and a new CI step that runs the full test suite on Windows.
- ›Provides better status messages explaining why ctags is disabled when it cannot be used.
- ›Adds
- v0.7.0
Aider v0.7.0 adds
--messagefor one-shot GPT instructions and--no-streamfor token cost visibility.└──▷ GET THIS VERSION$ git clone --branch v0.7.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.7.0
└──▷ TRY ITRun a one-shot refactor instruction in CI without an interactive session.$ aider --message 'Refactor all print statements to use logging' src/app.py
Disable streaming to capture token usage and cost info after a GPT edit session.$ aider --no-stream src/app.py
- ›Adds
--messageflag to send a single instruction to GPT and exit after it replies and applies any edits — enabling non-interactive, scriptable use. - ›Adds
--no-streamflag to disable streaming GPT responses; non-streaming mode surfaces token usage info and displays cost estimates based on OpenAI advertised pricing. - ›New coding competence benchmarking tool for evaluating GPT performance against a suite of programming tasks drawn from the Exercism Python repository.
- ›Initial function-based code editing backend for GPT-3.5, laying groundwork for OpenAI function calls API support.
- ›Adds graceful handling of context window exhaustion with user-facing tips when the limit is hit.
- ›Adds
- v0.6.2
Aider v0.6.2 adds
--openai-api-basefor API proxy support and extends model coverage to all OpenAI chat models includinggpt-3.5-turbo-16k.└──▷ GET THIS VERSION$ git clone --branch v0.6.2 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.6.2
└──▷ TRY ITRoute Aider through a local OpenAI-compatible proxy (e.g. LiteLLM, LocalAI) instead ofapi.openai.com.$ aider --openai-api-base http://localhost:8080/v1 --model gpt-3.5-turbo-16k
- ›Adds
--openai-api-baseCLI flag to point Aider at an OpenAI-compatible API proxy or alternative endpoint. - ›Adds support for
gpt-3.5-turbo-16kand all OpenAI chat models, expanding the selectable model surface.
- ›Adds
- v0.5.0
Aider v0.5.0 adds GPT-3.5/4-32k support, token budgeting with
--map-tokens, and new/tokensand/clearchat commands.└──▷ GET THIS VERSION$ git clone --branch v0.5.0 https://github.com/Aider-AI/aider.git # already have the repo? check out this version: $ git checkout v0.5.0
└──▷ TRY ITCap the repo map to a specific token budget so large codebases fit within context limits when using GPT-4-32k.$ aider --map-tokens 4096 --model gpt-4-32k
Check how many tokens the current conversation is consuming before adding more context.$ /tokens- ›Adds
--map-tokensCLI flag to set a token budget for the repo map, using a PageRank-based algorithm to prioritize which files and identifiers are included. - ›Adds in-chat command
/tokensto report current context window token usage. - ›Adds in-chat command
/clearto clear the conversation history. - ›Adds support for
gpt-3.5-turboandgpt-4-32kmodels.
- ›Adds