Heads up This site is currently under heavy development.
Subscribe Get it delivered — the daily firehose, filtered to the tools you run, plus the documentation changes vendors never announce. Compare plans →

The AI Toolchain — issue -361, August 20, 2025

THE AI TOOLCHAIN NO. -361
Tail
THE DAILY RELEASE FIREHOSE
PUBLISHED AUGUST 20, 2025 · EVERY WEEKDAY
EDITIONS tail grep head diff uniq

The daily firehose — everything the toolchain shipped today, already filtered.

// HOW THIS ISSUE IS MADE

We read every release from the 174 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.

VIEW
ISSUE VIEW full issue
Do you prefer this view?
$ tct list   # 14 tools matched
AI & LLM Tooling
◆  AI Agent Frameworks

Agno (formerly Phidata)

Sources Release notes → v1.7.12 NOTES

Agno v1.7.12 adds Team collaborate streaming, OpenAI verbosity control, Gemini URL context tool, Vertex AI embedder support, and DynamoDB throughput config.

└──▷ GET THIS VERSION
$ git clone --branch v1.7.12 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v1.7.12
└──▷ USE IT
Tune how much detail OpenAI returns in responses by setting verbosity on OpenAIChat.
python
from agno.models.openai import OpenAIChat

model = OpenAIChat(id="gpt-4o", verbosity=2)
  • Adds verbosity parameter to OpenAIChat and OpenAIResponses for controlling OpenAI output verbosity.
  • Allows specifying DynamoDB provisioned throughput when initializing DynamoDbStorage, enabling capacity control at construction time.
  • Adds Vertex AI support for GeminiEmbedder, enabling Gemini embeddings via the Vertex AI backend.
  • Adds streaming support for Team in collaborate mode, delivering incremental output for multi-agent collaborative workflows.
  • Adds support for the Gemini URL context tool, enabling Gemini models to fetch and reason over web URLs as context.
Was this useful?

AutoGPT

Sources Release notes → autogpt-platform-beta-v0.6.23 NOTES

AutoGPT Platform adds Enrichlayer and v0 by Vercel integrations, LaunchDarkly feature flags, and editing for agent submissions.

└──▷ GET THIS VERSION
$ git clone --branch autogpt-platform-beta-v0.6.23 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:
$ git checkout autogpt-platform-beta-v0.6.23
  • Adds support for v0 by Vercel models and credentials as a new LLM provider integration.
  • Adds Enrichlayer integration for data enrichment workflows.
  • Implements LaunchDarkly feature flag support for the block menu redesign.
  • Adds frontend service to docker-compose with environment config improvements.
  • Streamlines Supabase to minimal essential services in the platform stack.
+4 moreshow less
  • Enables editing for agent submissions.
  • New library agent view setup.
  • Adds a Line Tabs UI component.
  • Adds a reusable component for the new block menu.
Was this useful?

Letta (formerly MemGPT)

Sources Release notes → 0.11.4 NOTES

Letta 0.11.4 adds step metrics recording and asyncio stream-timeout protection.

└──▷ GET THIS VERSION
$ git clone --branch 0.11.4 https://github.com/letta-ai/letta.git
# already have the repo? check out this version:
$ git checkout 0.11.4
  • Records step metrics to a database table, enabling persistent tracking of per-step performance data.
  • Introduces asyncio shield around streaming calls to prevent stream timeouts under async workloads.
└──▷ BREAKING ON UPGRADE
  • !Legacy provider paths for Azure and Together AI are deprecated and may break configurations that relied on those routes.
Was this useful?

PydanticAI

Sources Release notes → v0.7.4 NOTES

PydanticAI v0.7.4 adds takes_ctx to Tool.from_schema and supports Google's url_context built-in tool.

└──▷ GET THIS VERSION
$ git clone --branch v0.7.4 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.7.4
└──▷ USE IT
Create a schema-derived tool that receives the agent context, enabling context-aware logic inside the tool handler.
python
tool = Tool.from_schema(schema=my_schema, takes_ctx=True)
Enable Google's URL context built-in tool so the agent can fetch and reason over live web content during a run.
python
from pydantic_ai.models.google import UrlContextTool

agent = Agent(model='google-gla:gemini-2.0-flash', tools=[UrlContextTool()])
  • Adds takes_ctx argument to Tool.from_schema, letting callers control whether the generated tool receives the agent context.
  • Supports Google's url_context built-in tool via the new UrlContextTool class, now exported in __all__.
Was this useful?

Microsoft Semantic Kernel

Sources Release notes → dotnet-1.63.0 NOTES

Semantic Kernel .NET 1.63.0 adds Guid/ObjectId key support for MongoDB and enriches OpenAPI parameter schemas.

└──▷ GET THIS VERSION
$ git clone --branch dotnet-1.63.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout dotnet-1.63.0
  • Adds support for Guid and ObjectId key types in the MongoDB Connector, expanding the range of document ID formats usable with the vector store integration.
  • Adds OpenAPI parameter descriptions to their JSON schemas, improving function metadata surfaced to the kernel when importing OpenAPI plugins.
Was this useful?

holmesgpt

Sources Release notes → 0.13.0 NOTES

SRE Agent - CNCF Sandbox Project

HolmesGPT 0.13.0 adds a Datadog RDS analysis toolset, REASONING_EFFORT env var, streaming API chat, custom annotations, and subtask support.

└──▷ GET THIS VERSION
$ git clone --branch 0.13.0 https://github.com/HolmesGPT/holmesgpt.git
# already have the repo? check out this version:
$ git checkout 0.13.0
  • Adds support for the REASONING_EFFORT environment variable to control LLM reasoning effort at runtime.
  • Adds call_stream support for the API chat endpoint, enabling streaming responses from the Holmes API.
  • Adds an option for custom annotations on Holmes resources.
  • Adds a new Datadog RDS analysis toolset for database-focused investigations via Datadog.
  • Adds subtask support, allowing Holmes investigations to spawn and track sub-investigations.
+7 moreshow less
  • Adds intermediate AI message and reasoning output via ai_message in investigation responses.
  • Adds LLM support to the investigation workflow, broadening which models can drive investigations.
  • Adds Prometheus AMP (Amazon Managed Prometheus) support alongside fixes to loading configs from environment variables.
  • Adds icon_url field to tool call results in the API response.
  • Adds revisionhistorylimit support for Kubernetes deployment configuration.
  • Returns rate-limit errors to callers when the upstream LLM rate-limits Holmes.
  • Adds an AI safety prompt to the system prompt for all Holmes interactions.
Was this useful?
◆  AI Coding Agents

Cline

Sources Release notes → v3.26.0 NOTES

Autonomous coding agent as an SDK, IDE extension, or CLI assistant.

Cline v3.26.0 adds Z AI GLM-4.5 models, LM Studio v0 API support, cline/sonic stealth model, and enables focus chain by default.

└──▷ GET THIS VERSION
$ git clone --branch v3.26.0 https://github.com/cline/cline.git
# already have the repo? check out this version:
$ git checkout v3.26.0
  • Adds support for Z AI GLM-4.5 and GLM-4.5 Air models.
  • Supports LM Studio local models via the v0 API endpoint.
  • Adds the cline/sonic stealth model.
  • Enables focus chain by default.
Was this useful?

Continue

Sources Release notes → v1.0.24-vscode NOTES

Continue v1.0.24 adds GPT-5, LGAI EXAONE support, MCP OAuth, and org policy enforcement.

└──▷ GET THIS VERSION
$ git clone --branch v1.0.24-vscode https://github.com/continuedev/continue.git
# already have the repo? check out this version:
$ git checkout v1.0.24-vscode
└──▷ USE IT
Enable autocomplete tuning in a YAML config file now that autocompleteOptions is supported there.
yaml
autocompleteOptions:
  debounceDelay: 350
  maxPromptTokens: 1024
  • Adds autocompleteOptions support in .yaml config files for controlling autocomplete behavior.
  • Adds MCP OAuth authentication support.
  • Adds GPT-5 model support.
  • Adds support for LGAI EXAONE models.
  • Enforces org policies, allowing organizations to control user-facing configuration options.
+4 moreshow less
  • Adds more next edit endpoints to the protocol.
  • Adds a tool call status icon for the search and replace UI.
  • Adds more prompt templates.
  • Opens conversation summary by default.
Was this useful?

Charm Crush

Sources Release notes → v0.7.0 NOTES

Glamourous agentic coding for all

Crush v0.7.0 adds a --data-dir flag and git-style parent-directory discovery for the .crush state folder.

└──▷ GET THIS VERSION
$ git clone --branch v0.7.0 https://github.com/charmbracelet/crush.git
# already have the repo? check out this version:
$ git checkout v0.7.0
└──▷ TRY IT
Point Crush at a shared data directory on a per-invocation basis, useful when you want to keep project state outside the default location.
$ crush --data-dir /path/to/project/.crush
  • Adds --data-dir (-D) flag to specify a custom path for the .crush data directory at runtime.
  • Automatically discovers a .crush directory in any parent directory, similar to how git locates .git, so subdirectory work reuses the project's data dir.
Was this useful?

OpenAI Codex CLI

Sources Release notes → rust-v0.23.0 NOTES

Lightweight coding agent that runs in your terminal

Codex CLI gains /mcp, /approvals, and /model slash commands plus runtime reasoning effort control and ChatGPT auth auto-refresh.

└──▷ GET THIS VERSION
$ git clone --branch rust-v0.23.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.23.0
└──▷ TRY IT
Switch model and tune reasoning effort mid-session without restarting — useful when a task shifts from quick edits to deep reasoning.
$ /model
Lock down or loosen what Codex is allowed to do automatically during a sensitive refactor.
$ /approvals
List or interact with connected MCP servers directly from the TUI.
$ /mcp
  • Adds /mcp slash command in the TUI to inspect and interact with MCP servers.
  • Adds /approvals slash command to control approval/permission settings at runtime.
  • Adds /model slash command to change the active model and reasoning effort without restarting.
  • Adds 'minimal' reasoning effort level for GPT-5 models.
  • Shows login options when not signed in with ChatGPT, improving auth discovery.
+7 moreshow less
  • Auto-refreshes ChatGPT auth tokens to maintain uninterrupted sessions.
  • Shows a progress indicator for the /diff command.
  • Adds a /diff command for diffing session changes.
  • Releases zip-archived binaries for easier distribution.
  • Adds DotSlash entry for Windows x86_64, enabling Windows distribution support.
  • Adds Ctrl+H as an additional backspace keyboard shortcut in the TUI.
  • Supports Ghostty terminal Ctrl-b/Ctrl-f word-navigation fallback in the TUI.
Was this useful?

All Hands AI OpenHands

Sources Release notes → 0.54.0 NOTES

OpenHands: AI-Driven Development

OpenHands 0.54 adds task tracking with planning mode, MCP on Cloud, and an improved MCP config UI.

└──▷ GET THIS VERSION
$ git clone --branch 0.54.0 https://github.com/All-Hands-AI/OpenHands.git
# already have the repo? check out this version:
$ git checkout 0.54.0
└──▷ HOW TO FIND IT
Connect an MCP server on OpenHands Cloud to extend the agent with custom tools.
📍In the OpenHands Cloud console, go to Settings › MCP and use the Add, Edit, or Delete controls to manage your MCP server configurations.
  • Adds a task tracking tool and planning mode to help agents manage long-horizon, multi-step tasks.
  • Enables Model Context Protocol (MCP) support on OpenHands Cloud.
  • Improves the MCP configuration UI with add, edit, and delete functionality.
  • Adds a new user feedback reason: 'The agent didn't finish the job'.
Was this useful?

Alibaba Qwen Code

Sources Release notes → v0.0.8-nightly.5 NOTES

Qwen Code v0.0.8-nightly.5 adds .geminiignore support in grep and a project/global save location option.

└──▷ GET THIS VERSION
$ git clone --branch v0.0.8-nightly.5 https://github.com/QwenLM/qwen-code.git
# already have the repo? check out this version:
$ git checkout v0.0.8-nightly.5
└──▷ USE IT
Exclude build artifacts and secrets from grep results by listing them in .geminiignore at the repo root.
yaml
# .geminiignore
dist/
*.env
node_modules/
*.log
  • Adds .geminiignore support to the grep tool, so files and paths listed in .geminiignore are excluded from grep searches.
  • Adds a project/global save location option, letting users choose whether settings are saved at the project or global scope.
Was this useful?

Zed

Sources Release notes → v0.200.4 NOTES

Zed v0.200.4 adds OS-specific settings, Git clone, OpenAI reasoning effort control, Mistral thinking, and file diff from the project panel.

└──▷ GET THIS VERSION
$ git clone --branch v0.200.4 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.200.4
└──▷ USE IT
Set macOS-specific font size without affecting your Windows or Linux Zed config.
json
{
  "macos": {
    "buffer_font_size": 14
  },
  "windows": {
    "buffer_font_size": 13
  }
}
Tune reasoning effort for a custom OpenAI model entry to get more thorough (or faster) responses.
json
{
  "language_models": {
    "openai": {
      "available_models": [
        {
          "name": "o3",
          "display_name": "o3 (high reasoning)",
          "max_output_tokens": 128000,
          "max_tokens": 200000,
          "reasoning_effort": "high"
        }
      ],
      "version": "1"
    }
  }
}
Keep the project panel closed by default when opening new projects, reducing visual noise.
json
{
  "project_panel": {
    "starts_open": false
  }
}
  • Adds top-level "macos", "windows", and "linux" fields in user settings to configure OS-specific overrides, with lower precedence than release-channel settings.
  • Adds reasoning_effort field (values: minimal, low, medium (default), high) to OpenAI model entries under language_models.openai.available_models for controlling reasoning effort per model.
  • Adds project_panel.starts_open setting to control whether the project panel opens automatically in new projects.
  • Adds status_bar.show_active_language_button setting to show or hide the language button in the status bar.
  • Adds git.inline_blame.padding setting for configurable padding on inline blame annotations.
+15 moreshow less
  • Adds theme_overrides setting for theme-specific setting overrides.
  • Adds git: Clone action to clone remote Git repositories from within Zed.
  • Adds compare marked files context menu item in the project panel to diff any two files against each other.
  • Adds Subtle and Eager edit prediction modes for Copilot and Supermaven.
  • Adds thinking support for the Mistral provider in the Agent.
  • Adds support for Ollama gpt-oss model in the Agent.
  • Adds editor: unwrap syntax node action for the editor.
  • Adds keymap contexts for pickers: CommandPalette, GitBranchSelector, GitRepositorySelector, RecentProjects, LanguageSelector, IconThemeSelector, ThemeSelector.
  • Agent now creates checkpoints before every edit, not just periodically.
  • OpenAI-compatible models now have configurable capabilities.
  • Adds Run to Cursor back to the editor context menu in the Debugger.
  • Go: Adds support for running Go table-test subtests.
  • Vue: Adds Emmet support in Vue.js files.
  • Vim: Adds filename support in :tabnew and :tabedit commands.
  • Vim: Adds ctrl-y / ctrl-e in insert mode to copy the next character from the line above or below.
└──▷ BREAKING ON UPGRADE
  • !FreeBSD support has been removed temporarily.
  • !The default split directions for pane: split horizontal and pane: split vertical have changed; restore the old behavior via pane_split_direction_horizontal and pane_split_direction_vertical in settings.
  • !always_allow_tool_actions is removed from project-local settings and is now a global-only setting.
  • !Settings overrides (e.g. local project settings) can now only override disable_ai to true; they can no longer re-enable AI that has been disabled at a higher level.
Was this useful?

Google gemini-cli

Sources Release notes → v0.2.0-preview.0 NOTES

An open-source AI agent that brings the power of Gemini directly into your terminal.

gemini-cli v0.2.0-preview.0 adds automatic conversation saving, Zed editor integration, MCP server restart, and tool auto-approval.

└──▷ GET THIS VERSION
$ git clone --branch v0.2.0-preview.0 https://github.com/google-gemini/gemini-cli.git
# already have the repo? check out this version:
$ git checkout v0.2.0-preview.0
└──▷ TRY IT
Restart a misbehaving MCP server mid-session without quitting gemini-cli.
$ /mcp refresh
  • Introduces ChatRecordingService for automatic conversation saving to session files.
  • Adds read/write file access through the Zed editor via IDE integration.
  • Adds hideFooter setting to hide the footer from the CLI UI.
  • Enables /mcp refresh to restart MCP servers in addition to refreshing them.
  • Shows /ide subcommands dynamically based on actual IDE connection status rather than a static boolean.
+5 moreshow less
  • Adds IDE client information to /bug and /about output when IDE mode is enabled.
  • Adds skipNextSpeakerCheck setting to control speaker-check behavior in core.
  • Auto-approves compatible pending tools when 'allow always' is selected, reducing repetitive prompts.
  • Refactors shell execution to use node-pty for improved terminal emulation support.
  • Adds file change tracking metrics to session telemetry.
Was this useful?
my-toolchain — 0 tools
paste an install list to detect your tools

A brew list, a Brewfile, requirements.txt, a Dockerfile — or just the product names, free-form. Nothing leaves your browser.

    browse all tools →