Backfill issue -315 — The AI Toolchain

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   # 8 tools matched

AI & LLM Tooling

◆ AI Coding Agents

Continue

@continuedev/[email protected] 2 RELEASES · 2025-10-06 NOTES STABLE

Continue 1.26.0 adds workflow files for CLI, API-key auth for Bedrock, and OpenRouter with Anthropic prompt caching.

└──▷ GET THIS VERSION

$ git clone --branch @continuedev/[email protected] https://github.com/continuedev/continue.git
# already have the repo? check out this version:
$ git checkout @continuedev/[email protected]

›Adds workflow file support for the CLI, including parsing, serializing, and tool permissions driven by workflow files.
›Adds API key authentication support for the Bedrock provider.
›Adds OpenRouter API integration with Anthropic prompt caching support.

1 more release in this issue · 2025-10-06

@continuedev/[email protected] NOTES STABLE

Continue's OpenAI adapters v1.23.0 adds API key authentication support for the Bedrock provider.

└──▷ GET THIS VERSION

$ git clone --branch @continuedev/[email protected] https://github.com/continuedev/continue.git
# already have the repo? check out this version:
$ git checkout @continuedev/[email protected]

›Adds API key authentication support for the Bedrock provider, enabling Bedrock access without IAM role or credential-file configuration.

Was this useful?

GitHub Copilot CLI

v0.0.335 NOTES

GitHub Copilot CLI v0.0.335 adds inline file diffs in the timeline and slash command argument hints.

└──▷ GET THIS VERSION

$ git clone --branch v0.0.335 https://github.com/github/copilot-cli.git
# already have the repo? check out this version:
$ git checkout v0.0.335

›Shows file diffs in the timeline by default, giving immediate visibility into file edits without requiring Ctrl+R.
›Displays argument hints inline in the slash command input box to guide command usage.
›Adds a warning when proxy support is used in an incompatible environment (Node <24 or required environment variables not set).
›Improves interface display in terminals narrower than 80 columns.
›Reduces color usage and improves spacing in Markdown rendering.

Was this useful?

OpenAI Codex CLI

rust-v0.45.0 NOTES

Lightweight coding agent that runs in your terminal

Codex CLI gains OAuth MCP auth, parallel tool calls, and gpt-5 default for Windows in v0.45.0

└──▷ GET THIS VERSION

$ git clone --branch rust-v0.45.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.45.0

└──▷ TRY IT

Securely log in to Codex CLI using an API key from your environment without exposing it on the command line.

$ printenv OPENAI_API_KEY | codex login --with-api-key

›Adds experimental OAuth authentication support for MCP servers.
›Removes the requirement to set experimental_use_rmcp_client for streamable HTTP MCP servers that don't need OAuth.
›Adds initial parallel tool call support.
›Defaults to gpt-5 for Windows users who have not configured a model.
›Outputs only the final message to stdout for codex exec, enabling cleaner scripting pipelines.

+1 moreshow less

›Adds a pulsing dot loading state indicator in the UI.

└──▷ BREAKING ON UPGRADE

!The codex login --api-key flag is replaced by --with-api-key, which reads the key from stdin; existing scripts passing the key directly via --api-key will break.

Was this useful?

Zed

v0.206.7 NOTES

Zed v0.206.7 adds Grok 4 Fast models and xAI custom model tool/image configuration.

└──▷ GET THIS VERSION

$ git clone --branch v0.206.7 https://github.com/zed-industries/zed.git
# already have the repo? check out this version:
$ git checkout v0.206.7

›Adds support for configuring tool and image support for xAI custom model configurations.
›Adds support for Grok 4 Fast models.

Was this useful?

◆ AI Agent Frameworks

LangChain

langchain-core==1.0.0a7 3 RELEASES · 2025-10-06 NOTES STABLE

langchain-core 1.0.0a7 adds PDF tool messages, optional include_id in OpenAI message conversion, Bedrock document blocks, and custom Mermaid URLs.

└──▷ GET THIS VERSION

$ git clone --branch langchain-core==1.0.0a7 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-core==1.0.0a7

└──▷ USE IT

Include message IDs when converting chat history to OpenAI format, useful for correlating messages across systems.

python

from langchain_core.messages.utils import convert_to_openai_messages
openai_msgs = convert_to_openai_messages(messages, include_id=True)

Strip PostgreSQL NUL bytes from LLM output before inserting into a database to avoid DataError.

python

from langchain_core.utils import sanitize_for_postgres
safe_text = sanitize_for_postgres(llm_output)

›Adds optional include_id parameter to convert_to_openai_messages function to control whether message IDs are included in OpenAI-format output.
›Adds id field to Document objects passed to the filter callback in InMemoryVectorStore similarity search.
›Adds web_search to the list of recognized OpenAI built-in tools.
›Adds image_generation to the list of recognized OpenAI built-in tools.
›Adds sanitize_for_postgres utility function to strip PostgreSQL NUL bytes that cause DataError.

+13 moreshow less

›Adds ls_model_name override support via kwargs on model invocations.
›Adds support for AWS Bedrock document content blocks in msg_content_output.
›Adds support for PDF inputs in ToolMessage content (shared with standard-tests).
›Adds support for PromptTemplate formats other than f-string.
›Adds Union type argument support in strict mode for OpenAI function calling and structured output.
›Adds an option to make deserialization more permissive.
›Adds additional hashing options to the indexing API and warns when SHA-1 is used.
›Supports custom Mermaid diagram URL via allow custom Mermaid URL capability.
›Exposes recognized block types for ToolMessage content.
›Traces response body on error in LangChain tracing.
›Zeroes out token costs for cache hits in token usage tracking.
›Batches Incremental record manager deletion for improved scalability.
›Removes Python upper bound restriction for langchain and co-library packaging.

2 more releases in this issue · 2025-10-06

langchain==1.0.0a12 NOTES STABLE

LangChain v1 alpha adds middleware hooks, PII/tool-call-limit/fallback middleware, async agent support, and model call limits.

└──▷ GET THIS VERSION

$ git clone --branch langchain==1.0.0a12 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain==1.0.0a12

›Adds before_agent and after_agent hooks to the langchain_v1 agent lifecycle for pre- and post-processing.
›Introduces retry_model_request middleware hook and ModelFallbackMiddleware for automatic model fallback on failure.
›Adds ToolCallLimitMiddleware to cap the number of tool calls an agent can make in a session.
›Implements PIIMiddleware for PII detection and handling in the agent pipeline.
›Adds LLM selection middleware, enabling dynamic routing of requests to different models.

+14 moreshow less

›Introduces Context Editing Middleware for modifying context mid-conversation.
›Adds async support to create_agent, enabling fully async agent construction and execution.
›Adds middleware support inside create_agent for composable, reusable agent behavior.
›Implements a dynamic system prompt middleware for runtime prompt modification.
›Adds a decorator pattern for dynamically generated middleware.
›Adds model call limits to langchain for controlling per-session or per-request model usage.
›Supports PEP 604 (| union) syntax in tool node error handlers.
›Enables stream_usage by default in the OpenAI integration when using the default base URL and client.
›Adds improvements to Anthropic prompt caching support.
›Introduces a description generator for Human-in-the-Loop (HITL) middleware.
›Adds improved HITL patterns with updated interrupt handling.
›Adds stuff and map reduce chains to langchain.
›Adds minimal and verbosity options to the OpenAI integration.
›Represents server-side tools in modifyModelRequest with updated tool handling.

└──▷ BREAKING ON UPGRADE

!The model_request node is renamed to model in langchain_v1.
!ToolNode support is removed from create_agent in langchain_v1.
!Text splitters are removed from the langchain_v1 namespace.
!Global state is removed from langchain-v1; globals in langchain-classic and langchain-core are updated.
!Python 3.9 is no longer supported in langchain v1.

langchain==1.0.0a11 NOTES STABLE

LangChain 1.0.0a11 adds middleware hooks, PII/fallback/tool-call-limit middleware, async agent support, and model call limits.

└──▷ GET THIS VERSION

$ git clone --branch langchain==1.0.0a11 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain==1.0.0a11

›Adds before_agent and after_agent lifecycle hooks to instrument or modify agent execution at entry and exit points.
›Adds retry_model_request middleware hook and ModelFallbackMiddleware to automatically retry or fall back to alternate models on failure.
›Adds ToolCallLimitMiddleware to cap the number of tool calls an agent can make in a single run.
›Adds PIIMiddleware to detect and redact personally identifiable information in model inputs or outputs.
›Adds LLM selection middleware, enabling dynamic routing of requests to different language models at runtime.

+17 moreshow less

›Adds Context Editing Middleware for modifying the context window passed to the model.
›Adds async support for create_agent, enabling non-blocking agent invocations.
›Adds model call limits to the langchain package, capping total model invocations.
›Adds middleware support inside create_agent, allowing middleware to be composed directly into agent construction.
›Adds dynamic system prompt middleware for runtime-generated system prompts.
›Adds a decorator pattern for dynamically generated middleware.
›Supports PEP 604 (| union) syntax in tool node error handlers.
›Adds stuff and map reduce chains to the library.
›Exposes rate_limiters from langchain_core in the langchain_v1 namespace.
›Represents server-side tools in modifyModelRequest and updates tool handling accordingly.
›Adds a description generator for Human-in-the-Loop (HITL) middleware.
›Improves HITL patterns with a structured response output schema key for the middleware agent.
›Adds improved Anthropic prompt caching support.
›Enables stream_usage by default when using the default base URL and client for the OpenAI integration.
›Adds minimal and verbosity options to the OpenAI integration.
›Adds todo middleware for tracking deferred actions within agent workflows.
›Adds a nicer developer experience for dynamic prompt construction.

└──▷ BREAKING ON UPGRADE

!ToolNode is removed from create_agent — setups passing ToolNode to create_agent will break.
!Text splitters are removed from the langchain_v1 namespace.
!Globals are removed from langchain-v1; globals in langchain-classic and langchain-core are updated — code relying on the old global locations will break.
!Python 3.9 is no longer supported.

Was this useful?

◆ Local LLM Runtimes

SGLang

v0.5.3 NOTES

SGLang v0.5.3 adds DeepSeek-V3.2 sparse attention, FlashAttention 4 prefill kernels, gRPC scheduler, and new model support.

└──▷ GET THIS VERSION

$ git clone --branch v0.5.3 https://github.com/sgl-project/sglang.git
# already have the repo? check out this version:
$ git checkout v0.5.3

›Adds a standalone gRPC server for the SGLang Python Scheduler, enabling direct gRPC-based scheduling workloads.
›Adds profile args support in the Engine API, exposing profiling controls programmatically.
›Adds polling interval configuration for prefill-decode disaggregation decode stage.
›Adds customer labels specified via request header to metrics, enabling per-request metric segmentation.
›Adds rerank routing logic in the regular router.

+19 moreshow less

›Adds embedding routing logic in the router.
›Allows one router to support different model families and serving modes simultaneously.
›Adds Basic OAI Response API to the router, along with get and cancel methods for the response API.
›Integrates FlashAttention 4 prefill kernels for accelerated attention computation.
›Supports loading model weights from a remote instance.
›Supports offloading in fp8.
›Supports memory_pool_host page-first direct layout.
›Adds Dynamic Batch Tokenizer for improved prefill throughput.
›Adds Generative Score API prefill-only scoring optimizations.
›Adds fine-grained request latency tracing (SGLang Tracing, Part 1).
›Adds latency histogram metrics for each stage of generate requests in PD disaggregation.
›Adds get_load watching mode and shortest-queue strategy to DP refactor.
›Adds split tile size option for Triton attention.
›Adds RTX 5880 MoE Triton kernel config.
›Supports global scale (in addition to per-expert scale) for CuteDSL MoE.
›Supports Qwen3-Next with MTP, DP, optimized kernels, and multiple hardware platforms including Ascend NPU and Blackwell.
›Supports new models: Qwen3-VL series, dots.vlm1, Ling-V2, Apertus, SOLAR.
›Adds ROCm Docker image support for sgl-router.
›Supports pip install sglang for simplified installation.

Was this useful?

Other / Uncategorized

◆ AI OBSERVABILITY

Arize Phoenix

arize-phoenix-v12.3.0 NOTES

Phoenix 12.3.0 adds dataset label management in the UI with a new settings tab and a label column in the datasets table.

└──▷ GET THIS VERSION

$ git clone --branch arize-phoenix-v12.3.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v12.3.0

›Adds a dataset label tab in the Settings page for managing dataset labels.
›Adds a label column to the datasets table for at-a-glance label visibility.
›Updates the experiment comparison slideover UI and adds pagination to the experiment comparison slideover on the list page.

Was this useful?

Langfuse

v3.116.0 NOTES

Langfuse v3.116.0 adds self-service enterprise plan, agent graph node cycling, and dataset compare view improvements.

└──▷ GET THIS VERSION

$ git clone --branch v3.116.0 https://github.com/langfuse/langfuse.git
# already have the repo? check out this version:
$ git checkout v3.116.0

›Adds self-service enterprise plan subscription option in billing UI.
›Enables cycling through observations in the agent graph by clicking individual graph nodes.
›Improves the dataset compare view UI to make cells more digestible.
›Expands content on hover in IO table cells for easier reading.

Was this useful?

// END OF TRANSMISSION

The AI Toolchain · curated daily · October 6, 2025
Releases & features from the open-source and commercial AI tools that practitioners actually run.

The AI Toolchain — issue -315, October 6, 2025