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.
$ git clone --branch langchain==1.2.11 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:$ git checkout langchain==1.2.11
›Adds langchain-openrouter provider package for integrating OpenRouter as a model provider.
›Supports automatic server-side compaction for OpenAI chat models.
LangGraph 1.1 adds opt-in version="v2" for type-safe streaming and invoke with Pydantic/dataclass output coercion.
└──▷ GET THIS VERSION
$ git clone --branch 1.1.0 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:$ git checkout 1.1.0
└──▷ USE IT
Get a typed return value and cleanly inspect interrupts after invoking a graph — no more fishing through result["__interrupt__"] in a plain dict.
python
result = graph.invoke({"input": "hello"}, version="v2")
result.value # your output state
result.interrupts # tuple[Interrupt, ...], empty if none
Stream graph events with full type narrowing — branch on part["type"] and let your type checker know exactly what part["data"] contains for each mode.
python
from langgraph.types import ValuesStreamPart, UpdatesStreamPart
for part in graph.stream({"input": "hello"}, version="v2"):
if part["type"] == "values":
state = part["data"] # OutputT — full typed state
interrupts = part["interrupts"]
elif part["type"] == "updates":
delta = part["data"] # dict[str, Any]
When your state is a Pydantic model, confirm the output is already coerced to the right type — no manual MyState(**result) call needed.
python
from pydantic import BaseModel
from langgraph.graph import StateGraph
class MyState(BaseModel):
answer: str
count: int
compiled = StateGraph(MyState).compile() # ... add nodes/edges first
result = compiled.invoke({"answer": "", "count": 0}, version="v2")
assert isinstance(result.value, MyState)
›Adds version="v2" opt-in to invoke(), ainvoke(), stream(), and astream() for fully type-safe outputs.
›New GraphOutput return type from invoke(..., version="v2") exposes .value and .interrupts attributes, cleanly separating state from interrupt signals.
›New strongly-typed StreamPart discriminated union (and per-mode TypedDicts: ValuesStreamPart, UpdatesStreamPart, MessagesStreamPart, CustomStreamPart, CheckpointStreamPart, TasksStreamPart, DebugStreamPart) enables full type narrowing in editors and type checkers.
›Automatic output coercion to Pydantic models or dataclasses when the graph's state schema is declared as one — no manual parsing needed.
›Non-"values" stream modes with version="v2" return list[StreamPart] from invoke() instead of list[tuple].
LangGraph CLI gains a langgraph deploy command for direct deployment from the CLI.
└──▷ GET THIS VERSION
$ git clone --branch cli==0.4.15 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:$ git checkout cli==0.4.15
└──▷ TRY IT
Deploy a LangGraph application to LangGraph Cloud without leaving the terminal.
$ langgraph deploy
›Adds langgraph deploy command to deploy LangGraph applications directly from the CLI.
LlamaIndex v0.14.16 adds token-bucket and sliding-window rate limiters, a multimodal reranker, GPT-5 and reasoning_content support, a ModelsLab LLM integration, and richer OpenTelemetry tracing.
└──▷ GET THIS VERSION
$ git clone --branch v0.14.16 https://github.com/run-llama/llama_index.git
# already have the repo? check out this version:$ git checkout v0.14.16
└──▷ USE IT
Use a custom embedding model inside the semantic double-merging splitter instead of the global default.
python
from llama_index.core.node_parser import SemanticDoubleMergingSplitterNodeParser
from llama_index.embeddings.openai import OpenAIEmbedding
parser = SemanticDoubleMergingSplitterNodeParser(
embed_model=OpenAIEmbedding(model="text-embedding-3-small")
)
Introspect the schema of a large Neo4j database by sampling with APOC rather than scanning all nodes.
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 v2.1.72 adds file-write in /copy, /plan descriptions, ExitWorktree tool, cron control, and a VS Code URI handler.
└──▷ GET THIS VERSION
$ git clone --branch v2.1.72 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:$ git checkout v2.1.72
└──▷ TRY IT
Write a focused selection to a file over SSH without relying on clipboard access.
$ # In an active Claude Code session, open the copy picker with /copy, highlight the desired block, then press `w` to write it directly to a file.
Jump straight into plan mode with context so Claude starts immediately rather than waiting for a follow-up prompt.
$ claude> /plan fix the auth bug
Stop all scheduled cron jobs mid-session without restarting Claude Code.
$ CLAUDE_CODE_DISABLE_CRON=1 claude
Open a new Claude Code tab in VS Code programmatically with a pre-filled prompt from an external script or tool.
$ open 'vscode://anthropic.claude-code/open?prompt=explain+this+file&session=abc123'
›Adds w key in /copy to write the focused selection directly to a file, bypassing the clipboard — useful over SSH.
›Adds optional description argument to /plan (e.g., /plan fix the auth bug) to enter plan mode and immediately start working.
›Adds ExitWorktree tool to cleanly leave an EnterWorktree session.
›Adds CLAUDE_CODE_DISABLE_CRON environment variable to immediately stop scheduled cron jobs mid-session.
›Adds lsof, pgrep, tput, ss, fd, and fdfind to the bash auto-approval allowlist, reducing permission prompts for common read-only operations.
+9 moreshow less
›Adds support for marketplace git URLs without a .git suffix (Azure DevOps, AWS CodeCommit).
›Adds claude plugins as an alias for claude plugin.
›Simplifies effort levels to low/medium/high (removed max) with new symbols (○ ◐ ●); adds /effort auto to reset to default.
›Improves /config keyboard UX — Escape cancels changes, Enter saves and closes, Space toggles settings.
›Improves up-arrow history to show the current session's messages first when running multiple concurrent sessions.
›Improves voice input transcription accuracy for repo names and common dev terms (regex, OAuth, JSON).
›Hides CLAUDE.md HTML comments (<!-- ... -->) from Claude when auto-injected; comments remain visible when read with the Read tool.
›VSCode: Adds vscode://anthropic.claude-code/open URI handler to open a new Claude Code tab programmatically, with optional prompt and session query parameters.
›VSCode: Adds effort level indicator on the input border.
└──▷ BREAKING ON UPGRADE
!The CLAUDE_CODE_PROXY_SUPPORTS_TOOL_REFERENCE environment variable has been removed; the tool search proxy bypass is now controlled by a different environment variable.
!Effort level max has been removed; valid levels are now low, medium, and high only.
Lightweight coding agent that runs in your terminal
Codex CLI gains runtime permission requests, a plugin marketplace, full web search config, and split sandbox policies.
└──▷ GET THIS VERSION
$ git clone --branch rust-v0.113.0 https://github.com/openai/codex.git
# already have the repo? check out this version:$ git checkout rust-v0.113.0
›Adds built-in request_permissions tool so a running turn can request additional permissions at runtime, with new TUI rendering for those approval prompts.
›Introduces a curated plugin marketplace with plugin/list metadata discovery, install-time auth checks, and a plugin/uninstall endpoint.
›Upgrades app-server command execution with streaming stdin/stdout/stderr and TTY/PTY support; wires exec to the new in-process app-server path.
›Expands web search tool configuration beyond on/off to support full options such as filters and location.
›Adds a new permission-profile config language and splits filesystem/network sandbox policies for more precise per-policy control.
+1 moreshow less
›Image generation now saves output files into the current working directory automatically.
llama.cpp server CORS proxy now correctly parses port numbers from MCP server URLs, enabling non-standard port and SSL routing.
└──▷ GET THIS VERSION
$ git clone --branch b8262 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:$ git checkout b8262
›The server CORS proxy now parses port numbers from MCP server URLs and passes the scheme to the HTTP proxy to determine whether to use SSL, fixing routing for non-standard ports.
llama.cpp b8261 extends Metal mul_mv_ext small-batch kernels to BF16, Q2_K, and Q3_K quantization types.
└──▷ GET THIS VERSION
$ git clone --branch b8261 https://github.com/ggml-org/llama.cpp.git
# already have the repo? check out this version:$ git checkout b8261
›Extends Metal mul_mv_ext small-batch kernels (batch sizes 2–8) to BF16, Q2_K, and Q3_K quantization types, which previously fell through to the slower single-row mul_mv path — enabling faster Apple Silicon inference for models using these formats.
Phoenix 13.13.0 adds provider filtering env vars, a DELETE session API, and dataset column drag-and-drop
└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v13.13.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:$ git checkout arize-phoenix-v13.13.0
└──▷ TRY IT
Restrict the playground and UI to only show specific model providers, e.g. to enforce org-approved vendors.
Weave v0.52.31 adds Claude agents integration, Fireworks provider support, trace server backend for tags and aliases, and timestamps/TTFT for realtime tracing.
└──▷ GET THIS VERSION
$ git clone --branch v0.52.31 https://github.com/wandb/weave.git
# already have the repo? check out this version:$ git checkout v0.52.31
›Adds Claude agents integration for tracing Anthropic agent workflows.
›Adds Fireworks as a supported LLM provider.
›Adds trace server backend support for tags and aliases on traced objects.
›Adds timestamps and time-to-first-token (TTFT) tracking for realtime tracing.
THE AI TOOLCHAIN