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.
AutoGPT v0.2.0 adds Selenium-based web browsing and a module-wrapped launcher for better extensibility.
└──▷ GET THIS VERSION
$ git clone --branch v0.2.0 https://github.com/Significant-Gravitas/AutoGPT.git
# already have the repo? check out this version:$ git checkout v0.2.0
›Adds Selenium-based web browsing capability, enabling AutoGPT to browse the web autonomously (requires Chrome installed and updated dependencies via pip install -r requirements.txt).
›Wraps AutoGPT in a module for launch, improving testability and extensibility of the runtime.
└──▷ BREAKING ON UPGRADE
!AutoGPT is now launched as a module; the startup method has changed — see the updated README.md for new run instructions.
!New dependencies (including Selenium) are required; existing installations must re-run pip install -r requirements.txt or the tool will not function correctly.
Haystack v1.16 adds GPT-4 and AzureChatGPT support, streaming, a Haystack CLI, and more flexible document routing.
└──▷ GET THIS VERSION
$ git clone --branch v1.16.0 https://github.com/deepset-ai/haystack.git
# already have the repo? check out this version:$ git checkout v1.16.0
└──▷ USE IT
Use GPT-4 in a multi-turn chat pipeline — drop-in for existing ChatGPT workflows with higher capability.
python
from haystack.nodes import PromptModel, PromptNode
prompt_model = PromptModel("gpt-4", api_key=api_key)
prompt_node = PromptNode(prompt_model)
messages = [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Summarize the attached document."},
]
result = prompt_node(messages)
›Adds PromptModel('gpt-4', api_key=...) support inside PromptNode and Agent, enabling chat-style multi-turn conversations with GPT-4.
›Adds AzureChatGPT invocation layer for PromptNode, enabling Azure-hosted ChatGPT endpoints via the new invocation layer style.
›Adds ChatGPT streaming support via PromptNode for real-time token-by-token output.
›Adds a Hugging Face Inference API invocation layer for PromptNode, enabling remote HF-hosted model inference without local GPU.
›Adds MemoryDocumentStore for the new Pipelines API.
+6 moreshow less
›Adds arbitrary crawler_depth parameter to the Crawler class, allowing configurable recursive web crawling depth.
›Enhances RouteDocuments node to emit an extra route for unmatched Documents and adds List[List[str]] support for metadata_values, preventing silent document loss on missing metadata fields.
›Adds filtering support for Weaviate when used for BM25 querying.
›Adds a Haystack CLI (haystack) for command-line management.
›Adds a load documents from remote helper function for fetching documents from remote sources.
›Deprecates RAGenerator and Seq2SeqGenerator; both will be removed in v1.18 — PromptNode is the recommended replacement.
└──▷ BREAKING ON UPGRADE
!Python 3.7 is no longer supported; upgrade to Python 3.8 or later.
!PreProcessor now requires farm-haystack[preprocessing]; installing the base package no longer pulls it in.
!DocxToTextConverter, TikaConverter, and LangdetectDocumentLanguageClassifier now require farm-haystack[file-conversion].
!ElasticsearchDocumentStore now requires farm-haystack[elasticsearch].
!TableCell replaces Span for indicating table cell coordinates.
!Default save_dir for FARMReader.train() changed to f'./saved_models/{self.inferencer.model.language_model.name}'.
!Using PreProcessor with split_respect_sentence_boundary=True may return a different set of Documents than in v1.15.
LangChain v0.0.150 adds DDG to load_tools, a Streamlit callback handler, PlugNPlai integration, ReAct eval chain, and Redis retriever document ingestion methods.
└──▷ GET THIS VERSION
$ git clone --branch v0.0.150 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout v0.0.150
└──▷ USE IT
Use DuckDuckGo search in an agent without needing an external API key, now that DDG is available via load_tools.
python
from langchain.agents import load_tools, initialize_agent
from langchain.llms import OpenAI
llm = OpenAI(temperature=0)
tools = load_tools(["ddg-search"], llm=llm)
agent = initialize_agent(tools, llm, agent="zero-shot-react-description", verbose=True)
agent.run("What is the latest news about LangChain?")
›Adds DDG (DuckDuckGo) as a supported tool in load_tools, enabling agent search without API keys.
›Adds add_documents and aadd_documents methods to RedisVectorStoreRetriever for synchronous and async document ingestion directly via the retriever class.
›Adds a Streamlit callback handler for streaming agent and chain output live into Streamlit apps.
›Adds PlugNPlai integration for loading and using plugins discovered via the PlugNPlai registry.
›Adds a ReAct eval chain for evaluating ReAct-style agent trajectories.
+3 moreshow less
›Adds a default request timeout for the Anthropic LLM integration.
›Adds Feast feature store integration notebook example.
›Adds Confluence loader with BeautifulSoup parsing support.
LangChain v0.0.143 adds eight new document loaders, a combining output parser, OpenSearch Boolean Filter support, and Redis/Jinja2 improvements.
└──▷ GET THIS VERSION
$ git clone --branch v0.0.143 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout v0.0.143
›Adds Redis.from_url() for initializing a Redis vector store directly from a connection URL.
›Adds support for Boolean Filter with ANN search in the OpenSearch integration, with kwargs passthrough to from_texts.
›Adds a shared ChromaDB client option, allowing multiple components to reuse a single chromadb.Client instance.
›Adds CombiningOutputParser to chain multiple output parsers together.
›Adds inference of input_variables from Jinja2 templates, so prompt templates no longer require manually listing variables when using the jinja2 template format.
+9 moreshow less
›Adds a GoogleSQL prompt for SQL chain integrations.
›Adds new document loader: Confluent (Kafka) loader.
›Adds new document loader: image caption loader.
›Adds new document loader: Jira loader.
›Adds new document loader: Twitter tweet loader.
›Adds new document loader: Obsidian loader.
›Adds new document loader: Discord loader.
›Updates CometML integration with new tracing capabilities.
›Updates HuggingFaceEmbeddings to support loading from cached weights.
Use Anthropic Claude as a drop-in chat model for a LangChain chain or agent.
python
from langchain.chat_models import ChatAnthropic
from langchain.schema import HumanMessage
chat = ChatAnthropic()
response = chat([HumanMessage(content="What are the top risks in a zero-trust architecture?")])
print(response.content)
›Adds openai.api_base parameter to OpenAI LLM to support routing through an OpenAI-compatible proxy.
›Adds GitLoader document loader with a file_filter parameter and automatic .gitignore exclusion for loading code repositories into LangChain.
›Adds ChatAnthropic chat model integration, bringing Anthropic's Claude models into the LangChain chat model interface.
›Adds Slack Directory Loader for ingesting Slack export directories as documents.
›Adds retriever-backed memory (Harrison/retriever memory), enabling chains to use vector retrieval for conversational context.
+6 moreshow less
›Adds dialect-specific prompts for SQLDatabaseChain, improving SQL generation accuracy across database backends.
›Supports PATCH and DELETE HTTP methods in reduce_openapi_spec, expanding OpenAPI chain coverage.
›Updates modelname_to_contextsize in the OpenAI LLM with new model context window sizes.
›Adds easy print method to the OpenAI callback handler for quick token usage inspection.
›Adds PyTorch 2 support for local model integrations.
›Adds Mendable Search integration as a retriever/tool.
LangChain v0.0.137 adds async APIChain, GPT4All streaming, PDF-as-HTML loading, OpenSearch custom fields, and an OpenAPI planner agent.
└──▷ GET THIS VERSION
$ git clone --branch v0.0.137 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout v0.0.137
└──▷ USE IT
Run an APIChain asynchronously inside an async application to avoid blocking the event loop.
python
import asyncio
from langchain.chains import APIChain
from langchain.llms import OpenAI
chain = APIChain.from_llm_and_api_docs(OpenAI(), api_docs='<your-api-docs>')
result = asyncio.run(chain.arun('What is the current weather in London?'))
print(result)
›Adds async support to APIChain via arun method, enabling non-blocking API chain calls.
›Adds streaming support for GPT4All LLM integration.
›Adds a new PDF loader that loads PDF content as HTML, expanding document ingestion options.
›Adds custom vector fields and text fields support for OpenSearch vector store.
›Adds special token params for tiktoken to OpenAIEmbeddings.
+5 moreshow less
›Adds a custom LLM option for the QueryChecker inside SqlDatabaseToolkit.
›Adds run and arun methods to document combination chains in place of combine_docs and acombine_docs.
›Adds a BabyAGI agent notebook example demonstrating autonomous task-management with LangChain.
›Adds a CAMEL role-playing multi-agent notebook example.
›Adds an OpenAPI planner agent for navigating and calling OpenAPI-described services.
└──▷ BREAKING ON UPGRADE
!combine_docs and acombine_docs are replaced by run and arun on document combination chains — any code calling combine_docs or acombine_docs directly will break.
LangChain v0.0.133 adds multi-action agents, an OpenAPI parser/spec toolkit, and Outlook email loading support.
└──▷ GET THIS VERSION
$ git clone --branch v0.0.133 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout v0.0.133
›Extends UnstructuredEmailLoader to support Microsoft Outlook files (.msg format) in addition to existing email formats.
›Introduces a multi-action agent that can emit and execute multiple tool actions in a single step, enabling more complex agentic workflows.
›Adds an OpenAPI parser and OpenAPI spec integration, enabling agents to interact with APIs described by an OpenAPI specification via a new agent toolkit.
KoboldCpp v1.16 adds Tail Free Sampling and Typical Sampling, plus CLBlast support for q5_0 and q5_1 formats.
└──▷ GET THIS VERSION
$ git clone --branch v1.16 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.16
›Adds CLBlast GPU acceleration support for the q5_0 and q5_1 quantization formats.
›Adds two new token samplers: Tail Free Sampling (TFS) and Typical Sampling, available alongside the reworked Top-P, Top-K, and Rep Pen samplers.
›Unifies sampling functions across all model architectures and types under a single overhauled sampling system.
└──▷ BREAKING ON UPGRADE
!Upstream llama.cpp has completely removed support for the q4_3 format; users are strongly advised to switch away from q4_3 and reconvert any existing q4_3 models.
16 more releases in this issue
· 2023-04-01 → 2023-04-30
›Adds --skiplauncher flag to bypass the new Easy Mode GUI and proceed directly to CLI operation.
›Adds --debugmode flag to print the tokenized prompt sent to the backend in the terminal window.
›Setting --stream now automatically redirects the embedded Kobold Lite UI to streaming mode, removing the need to manually append ?streaming=1 to the URL.
›Introduces a new Easy Mode GUI launcher that activates when no command-line arguments are provided, offering a guided setup for first-time users.
›Adds q5_0 and q5_1 quantization format support for llama.cpp, GPT-2, GPT-J, and GPT-NeoX model formats (OpenBLAS supported; CLBlast not yet supported).
+1 moreshow less
›Kobold Lite UI now supports multiple custom stopping sequences, separated by the ||$|| delimiter, with sequences saved to save files and autosaved.
KoboldCpp v1.11 adds GPT-NeoX/Pythia/StableLM support, --lora for llama, and multi-backend build improvements.
└──▷ GET THIS VERSION
$ git clone --branch v1.11 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.11
└──▷ TRY IT
Build both OpenBLAS and CLBlast backends on Linux/OSX, then select CLBlast at runtime for GPU acceleration.
$ make LLAMA_OPENBLAS=1 LLAMA_CLBLAST=1 && ./koboldcpp mymodel.bin --useclblast
›Adds --lora parameter to enable LORA file support for llama models.
›Adds GPT-NeoX, Pythia, and StableLM model architecture support.
›Adds limited fast-forwarding for RWKV, allowing context reuse when the context is completely unmodified.
›Kobold Lite UI now supports a custom stopping sequence, configurable in the Memory panel.
›Improved OSX and Linux builds now compile multiple acceleration backends (e.g. make LLAMA_OPENBLAS=1 LLAMA_CLBLAST=1) and allow selecting between them at runtime via flags such as --useclblast.
KoboldCpp v1.9 adds API stopping sequences support and BLAS mode for GPT-J and GPT2 models.
└──▷ GET THIS VERSION
$ git clone --branch v1.9 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.9
›Adds stopping sequences support to the API, allowing generation to halt early when a stop sequence is matched and return the response immediately without consuming remaining tokens.
›GPT-J and GPT2 models now support BLAS mode for faster inference, using a smaller batch size than LLaMA models.
$ git clone --branch v1.8.1 https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.8.1
›CLBlast now performs 4-bit dequantization on the GPU (via --useclblast [platform_id] [device_id]), delivering approximately 20% faster inference for CLBlast users.
KoboldCpp v1.0.9beta adds GPT-2 model support and Alpaca Instruct Mode in the embedded Kobold Lite UI.
└──▷ GET THIS VERSION
$ git clone --branch v1.0.9beta https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.0.9beta
›Adds GPT-2 model support (including theoretical compatibility with Cerebras models), enabling inference on very small ggml models at high token throughput on CPU.
›Adds Stanford Alpaca-compatible Instruct Mode to the embedded Kobold Lite interface, enabling structured prompt/response formatting — configurable in Kobold Lite settings.
›Adds repetition penalty (Rep Pen) support for GPT-J and GPT-2 models (and pyg.cpp), bringing penalty behavior in line with llama.cpp.
KoboldCpp v1.0.8beta adds GPT4ALL.CPP and GPT-J format support and boosts generation speed with -Ofast.
└──▷ GET THIS VERSION
$ git clone --branch v1.0.8beta https://github.com/LostRuins/koboldcpp.git
# already have the repo? check out this version:$ git checkout v1.0.8beta
└──▷ TRY IT
Override the new physical-core-based thread default to maximize throughput on a hyperthreaded or high-core-count machine.
$ koboldcpp.exe --threads 16 <model_path>
›Adds support for the original GPT4ALL.CPP model format.
›Adds support for GPT-J formats, including the original 16-bit legacy format and the 4-bit version from Pygmalion.cpp.
›Switches compiler optimization flag from -O3 to -Ofast, increasing token generation speed.
›Changes default thread count to scale by physical core count rather than os.cpu_count(), with manual override available via --threads.
└──▷ BREAKING ON UPGRADE
!Library file names and references are renamed as part of the rebranding from llamacpp-for-kobold to koboldcpp — any scripts or integrations referencing the old library names will break.
Triton v2.33.0 adds concurrent model loading, OpenTelemetry tracing, HTTP/gRPC header forwarding, configurable latency quantiles, and protocol access restrictions.
└──▷ GET THIS VERSION
$ git clone --branch v2.33.0 https://github.com/triton-inference-server/server.git
# already have the repo? check out this version:$ git checkout v2.33.0
›Adds experimental latency metrics as configurable quantiles over a sliding time window via metrics summary support (see metrics.md#summaries).
›Adds beta support for restricting access to specific protocols on a given Triton endpoint (see inference_protocols.md#limit-endpoint-access-beta).
›Adds experimental support for schedule policy in the sequence batcher with direct scheduling strategy.
›Adds limited support for tracing inference requests using OpenTelemetry Trace APIs.
›Enables forwarding of HTTP/gRPC headers as inference request parameters to the backend.
+4 moreshow less
›Extends ragged batching support to the PyTorch backend.
›Enables concurrent model loading to reduce server start-up times.
›Python backend business logic scripting (BLS) now allows selecting a specific device to receive output tensors from a BLS call.
›Model Analyzer adds support for BLS model config search.
shell-gpt 0.9.0 adds custom user-defined roles with --create-role, --list-roles, and --show-role flags.
└──▷ GET THIS VERSION
$ git clone --branch 0.9.0 https://github.com/TheR1D/shell_gpt.git
# already have the repo? check out this version:$ git checkout 0.9.0
└──▷ TRY IT
Create a reusable 'json' role so every prompt returns only valid JSON — useful for piping structured data into other tools.
$ sgpt --create-role json
# Enter role description: You are JSON generator, provide only valid json as response.
# Enter expecting result, e.g. answer, code, shell command, etc.: json
sgpt --role json "random: user, password, email, address"
›Adds --create-role <name> flag to define custom roles stored as JSON files in ~/.config/shell_gpt/roles, each specifying a system prompt and expected output type (answer, code, shell command, etc.).
›Adds --role <name> flag to invoke any custom or built-in role when running a prompt.
›Adds --list-roles flag to display all available roles, including user-created and built-in ones.
›Adds --show-role <name> flag to display the details of a specific role.
›Allows overriding the built-in shell, code, and default roles by editing their JSON files in ~/.config/shell_gpt/roles.
+2 moreshow less
›Adds option to force the use of system role messages via a dedicated flag (not recommended by the project).
›Improves stdin-plus-prompt handling, e.g. echo hello | sgpt "another hello".
└──▷ BREAKING ON UPGRADE
!All chats created with previous versions of ShellGPT are incompatible with 0.9.0 and will not work after upgrading.
!The --list-chat flag is renamed to --list-chats; any scripts or aliases using --list-chat will break.
4 more releases in this issue
· 2023-04-03 → 2023-04-16
shell-gpt 0.8.8 lets you combine stdin piping and a command-line prompt in a single invocation.
└──▷ GET THIS VERSION
$ git clone --branch 0.8.8 https://github.com/TheR1D/shell_gpt.git
# already have the repo? check out this version:$ git checkout 0.8.8
└──▷ TRY IT
Generate a git commit message by piping a diff into sgpt alongside an explicit prompt — no temp files needed.
$ git diff | sgpt "Generate git commit message, for my changes"
›Accepts a prompt from both stdin and a command-line argument simultaneously, enabling piped output to be combined with an inline instruction in one command.
shell-gpt 0.8.3 adds an interactive REPL mode for chat sessions via --repl, compatible with --shell and --code.
└──▷ GET THIS VERSION
$ git clone --branch 0.8.3 https://github.com/TheR1D/shell_gpt.git
# already have the repo? check out this version:$ git checkout 0.8.3
└──▷ TRY IT
Start an interactive shell-command session in REPL mode to iteratively build and refine commands without re-invoking sgpt each time.
$ sgpt --repl my-session --shell
Pick up an existing chat session inside REPL mode to continue a conversation with full history displayed.
$ sgpt --repl my-session
›Adds --repl <session-name> option to start an interactive REPL mode for chat sessions, showing conversation history on entry; accepts temp as a session name for a throwaway session.
›REPL mode shares sessions with --chat, allowing seamless hand-off between the two modes mid-conversation.
›REPL mode supports --shell and --code flags for interactive shell command generation and code generation within the same session.
›Adds prompt_column_names and response_column_names fields to Schema, each accepting EmbeddingColumnNames with vector_column_name and raw_data_column_name, enabling prompt/response pair ingestion for generative LLM workflows.
›Adds tag_column_names list field to Schema for attaching scalar metric columns (e.g. rouge scores) to LLM dataset entries.
›Adds prompt_column_names and response_column_names parameters to Schema, each accepting an EmbeddingColumnNames with vector_column_name and raw_data_column_name, to natively represent LLM prompt/response pairs in datasets.
›Renders grid previews of LLM prompts and responses in the embeddings UI.
›Displays prompt and response content in the event details panel for LLM inference events.
›Shows prompt/response pairs in the selection table and on inference event views.
Milvus v2.2.7 adds QueryNode plugin support for dynamic shared-library loading, replica-granularity load balancing, and a score-based balancing strategy.
└──▷ GET THIS VERSION
$ git clone --branch v2.2.7 https://github.com/milvus-io/milvus.git
# already have the repo? check out this version:$ git checkout v2.2.7
›Adds plugin logic to QueryNode to support dynamic loading of shared library files.
›Supports load balancing with replica granularity.
›Releases a score-based load-balancing strategy.
›Improves search grouping algorithm to enhance query throughput.
›Improves compaction algorithm to drive segment sizes toward an ideal distribution.
+3 moreshow less
›Adds a coroutine pool to limit concurrency of cgo calls triggered by delete operations.
›Reduces peak memory consumption during collection loading.
›Changes the default shard number to 1.
└──▷ BREAKING ON UPGRADE
!The default shard number is changed to 1; collections created without an explicit shard count will now have 1 shard instead of the previous default.
1 more release in this issue
· 2023-04-18 → 2023-04-28
Qdrant v1.1.1 adds per-vector HNSW/quantization config, TLS for gRPC and REST, isNull payload filter, and snapshot multipart upload.
└──▷ GET THIS VERSION
$ git clone --branch v1.1.1 https://github.com/qdrant/qdrant.git
# already have the repo? check out this version:$ git checkout v1.1.1
└──▷ TRY IT
Trigger snapshot creation without waiting for it to finish, so long-running snapshot jobs do not block your API call.
$ curl -X POST 'http://localhost:6333/collections/my_collection/snapshots?wait=false'
›Adds isNull condition for payload filtering, enabling queries that distinguish null values from empty or missing fields in specific payload keys.
›Adds wait parameter to the snapshot API, allowing callers to skip blocking on snapshot creation and return immediately — useful for long-running operations.
›Adds last-used and startup timing fields to the telemetry API response.
›Adds aggregated vector count to the /metrics endpoint.
›Adds per-vector-field HNSW and quantization configuration, so each named vector field in a collection can carry independent index and quantization settings.
+4 moreshow less
›Adds TLS support for gRPC and REST API, plus TLS for internal inter-node communication, with mutual (client and server) certificate verification.
›Adds ability to upload and recover snapshot files via multipart HTTP requests.
›Adds parameter validation to REST and gRPC APIs and to the config file, providing clearer error messages on misconfiguration.
›Introduces an internal rate limiter for the transport channel pool, improving cluster stability under high-concurrency load.
THE AI TOOLCHAIN