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.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
Claude Code v2.1.94 adds Mantle-powered Bedrock support, raises default effort to high, and expands hook/plugin capabilities.
└──▷ GET THIS VERSION
$ git clone --branch v2.1.94 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:$ git checkout v2.1.94
└──▷ TRY IT
Route Claude Code through Amazon Bedrock powered by Mantle instead of the default endpoint.
$ CLAUDE_CODE_USE_MANTLE=1 claude
Set a custom session title from a UserPromptSubmit hook by returning hookSpecificOutput.sessionTitle.
›Adds Amazon Bedrock powered by Mantle support via CLAUDE_CODE_USE_MANTLE=1.
›Raises default effort level from medium to high for API-key, Bedrock/Vertex/Foundry, Team, and Enterprise users (adjustable with /effort).
›Adds compact Slacked #channel header with a clickable channel link for Slack MCP send-message tool calls.
›Adds keep-coding-instructions frontmatter field support for plugin output styles.
›Adds hookSpecificOutput.sessionTitle to UserPromptSubmit hooks for setting the session title programmatically.
+3 moreshow less
›Plugin skills declared via "skills": ["./"] now use the skill frontmatter name as the invocation name instead of the directory basename, providing a stable name across install methods.
›Improves --resume to resume sessions from other worktrees of the same repo directly instead of printing a cd command.
›[VSCode] Adds a warning banner when settings.json files fail to parse so users know their permission rules are not being applied.
GitHub Copilot CLI v1.0.21 adds copilot mcp for MCP server management and snake_case hook payloads.
└──▷ GET THIS VERSION
$ git clone --branch v1.0.21 https://github.com/github/copilot-cli.git
# already have the repo? check out this version:$ git checkout v1.0.21
›Adds copilot mcp command for managing MCP servers.
›Hooks configured with PascalCase event names now receive VS Code-compatible snake_case payloads containing hook_event_name, session_id, and ISO 8601 timestamps.
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
CrewAI 1.14.0 adds runtime state checkpointing with SQLite storage, automatic checkpoint config, and new CLI commands for checkpoint management.
└──▷ GET THIS VERSION
$ git clone --branch 1.14.0 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:$ git checkout 1.14.0
└──▷ TRY IT
List all saved checkpoints for a crew run to find a resume point after an interrupted execution.
$ crewai checkpoint list
Inspect details of a specific checkpoint before deciding whether to resume from it.
$ crewai checkpoint info <checkpoint-id>
›Adds checkpoint list and checkpoint info CLI commands to inspect saved crew execution states.
›Introduces SqliteProvider for persistent checkpoint storage backend.
›Adds CheckpointConfig for configuring automatic checkpointing of crew runs.
›Implements runtime state checkpointing and an event system so long-running crews can be resumed after interruption.
›Adds guardrail_type and name fields to traces to distinguish guardrail events in observability pipelines.
+1 moreshow less
›Exports JsonProvider as an additional checkpoint storage option alongside SqliteProvider.
└──▷ BREAKING ON UPGRADE
!CodeInterpreterTool is removed and code execution parameters are deprecated — any crew or agent config referencing CodeInterpreterTool or those parameters will break on upgrade.
langchain-ollama 1.1.0 adds structured output, embedding dimensions, and logprobs support to Ollama integrations.
└──▷ GET THIS VERSION
$ git clone --branch langchain-ollama==1.1.0 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain-ollama==1.1.0
└──▷ USE IT
Request structured JSON output from an Ollama model in a typed workflow.
python
from langchain_ollama import ChatOllama
llm = ChatOllama(model="llama3", response_format={"type": "json_object"})
response = llm.invoke("Return a JSON object with keys 'host' and 'port' for a web server.")
print(response.content)
Generate fixed-size embeddings to match a downstream vector store's expected dimensionality.
Retrieve per-token log probabilities to assess model confidence in generated detections.
python
from langchain_ollama import ChatOllama
llm = ChatOllama(model="llama3", logprobs=True)
response = llm.invoke("Classify this log line as benign or malicious.")
print(response.response_metadata)
›Adds response_format parameter to ChatOllama for structured/JSON output control.
›Adds dimensions parameter to OllamaEmbeddings to specify output embedding vector size.
›Adds logprobs support to ChatOllama, enabling token-level log-probability output.
LangGraph SDK adds langsmith_tracing parameter to runs.create/stream/wait for per-call tracing control.
└──▷ GET THIS VERSION
$ git clone --branch sdk==0.3.13 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:$ git checkout sdk==0.3.13
└──▷ USE IT
Disable LangSmith tracing for a specific run invocation to avoid logging sensitive payloads.
$ git clone --branch b8685 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:$ git checkout b8685
›Extends the SYCL reorder optimization (scale-factor/weight separation for coalesced memory access) to Q8_0, achieving ~3.1x token-generation throughput on Intel Arc Pro B70 (Xe2) — e.g. 4.88 → 15.24 t/s on Qwen3.5-27B, with bandwidth utilization rising from 21% to 66%.
Connect a remote MCP server so its tools are automatically discovered and available alongside local tools during chat.
📍In the Chat tab, locate the 'MCP servers' field and enter your server URL (one per line), e.g.:
https://mcp.example.com/my-server
Then send a message — tools are discovered and invoked automatically.
›Adds instruction_template and instruction_template_str parameters to the model load API endpoint, allowing callers to specify prompt format at load time.
›Adds MCP server support: enter one remote MCP server URL per line in the new 'MCP servers' field in the Chat tab — tools are discovered automatically and used alongside local tools.
›PNG images returned by the image generation API now include generation metadata (model, seed, dimensions, steps, CFG scale, sampler) embedded in the file.
›Adds Windows + ROCm portable builds (ROCm 7.2) alongside the existing Windows CUDA and Vulkan builds.
›Updates ik_llama.cpp backend, adding Gemma 4 model support.
└──▷ BREAKING ON UPGRADE
!The deprecated settings parameter is removed from the model load endpoint — any client passing settings will need to be updated.
Phoenix v14 adds PostgreSQL read replicas, ephemeral experiments with ExperimentSweeper, a subcommand-first CLI, and Strands Agents integration.
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v14.0.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v14.0.0
└──▷ TRY IT
Start the Phoenix server in dev mode using the new subcommand-first CLI syntax.
$ phoenix serve --dev
›Restructures CLI into a subcommand-first architecture: flags now follow the subcommand (e.g. phoenix serve --dev, phoenix db ...) instead of preceding it.
›Adds PostgreSQL read replica routing support for load-distributing query traffic.
›Introduces ephemeral experiments and the ExperimentSweeper daemon for automated experiment lifecycle management.
›Adds a background experiment runner to execute experiment jobs asynchronously.
›Adds experiment record toggle with ephemeral tracking in the Playground UI.
+12 moreshow less
›Adds a structured error table with sanitized messages in the experiments view.
›Includes span and experiment run data in error subscription payloads.
›Polls the experiments table automatically while jobs are running.
›Adds Strands Agents integration with onboarding snippet.
›Adds new provider integrations and environment variable support to the onboarding flow.
›Adds ask_user elicitation tool with a carousel UI to the agent panel.
›Adds a backend MCP docs tool via Mintlify integration to the agent.
›Traces Phoenix agent chat requests both locally and remotely.
›Adds @defer support to Relay and ProjectPageHeader for progressive UI loading.
›Requires explicit first argument for forward pagination in the GraphQL API.
›Uses non-streaming mode for LLM evaluator calls.
›Uses raw vendor response as output.value for non-streaming Playground sessions.
└──▷ BREAKING ON UPGRADE
!The /v1/evaluations REST endpoint and its supporting plumbing have been removed.
!The legacy experiments module has been removed and evals 1.0 is deprecated.
!CLI flags must now follow the subcommand: phoenix serve --dev instead of phoenix --dev serve.
Langfuse v3.164.0 adds full-text search on input/output in table views, more fields in blob storage exports, and a new env var to exclude projects from experiment backfill.
└──▷ GET THIS VERSION
$ git clone --branch v3.164.0 https://github.com/langfuse/langfuse.git
# already have the repo? check out this version:$ git checkout v3.164.0
›Adds environment variable to exclude specific project IDs from experiment backfill processing (feat(worker): add env var to exclude project IDs from experiment backfill).
›Expands v3 and v4 blob storage exports with additional fields.
›Adds full-text search targeting input/output directly in table UI views.
›Adds common.threadCoreCoefficient.maxThreadsSize config key to make thread pool max thread count configurable with dynamic (no-restart) update support.
›Adds force-promote support for primary-secondary disaster-recovery failover via the UpdateReplicateConfiguration API.
›Adds data salvage capability for force failover, recovering unpersisted data from failed streaming nodes.
›Improves query filter performance with type-aware bidirectional rewriting between in and == expressions.
›Optimizes unfiltered search on sealed segments using MVCC fast path, hardware popcnt, and redundant bitset operation elimination.
+4 moreshow less
›Switches import retry strategy from allowlist to denylist, improving resilience against transient errors during data import.
›Reduces MixCoord recovery time by parallelizing startup phases including sub-meta loading, index reload, and batch etcd operations.
›Aligns error mapping across Azure, GCP, and MinIO for consistent retry and error-handling behavior.
›Optimizes bool IN/NOT IN expressions with proper nullable field handling.
THE AI TOOLCHAIN