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 -169, March 2, 2026

THE AI TOOLCHAIN NO. -169
Tail
THE DAILY RELEASE FIREHOSE
PUBLISHED MARCH 2, 2026 · 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   # 11 tools matched
AI & LLM Tooling
◆  AI Agent Frameworks

Agno (formerly Phidata)

Sources Release notes → v2.5.6 NOTES

Agno v2.5.6 adds GitHub App auth for knowledge sources, HEIC/HEIF uploads, approval endpoints, and advanced trace filtering DSL.

└──▷ GET THIS VERSION
$ git clone --branch v2.5.6 https://github.com/agno-agi/agno.git
# already have the repo? check out this version:
$ git checkout v2.5.6
└──▷ USE IT
Import Gmail tools from the new sub-package structure after the Google tools restructure.
python
from agno.tools.google import GmailTools

tools = GmailTools()
  • Adds GitHub App authentication to GitHubConfig via app_id, installation_id, and private_key parameters, with thread-safe token caching and both sync/async variants, in addition to existing personal access token support.
  • Adds image/heic and image/heif MIME type support to file upload endpoints.
  • Adds an approval status endpoint and admin-gated continue-run enforcement for agent workflows.
  • Adds advanced filtering DSL support for Traces in Agent OS.
  • Restructures Google tools into the agno.tools.google sub-package, enabling imports such as from agno.tools.google import GmailTools; old import paths remain functional via backwards compatibility.
+1 moreshow less
  • Adds tasks: List[TaskData] field (containing id, title, description, status, assignee, dependencies, result) and completion_summary to TaskStateUpdatedEvent for structured task data in TeamMode.tasks streaming.
Was this useful?

LangChain

Sources Release notes → langchain-huggingface==1.2.1 NOTES

langchain-huggingface 1.2.1 adds text_inputs and text_outputs fields to model profiles.

└──▷ GET THIS VERSION
$ git clone --branch langchain-huggingface==1.2.1 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-huggingface==1.2.1
  • Adds text_inputs and text_outputs fields to model profiles, enabling explicit declaration of text input/output surfaces per model.
Was this useful?

LangChain LangGraph

Sources Release notes → cli==0.4.14 NOTES

Build resilient agents.

LangGraph CLI gains a keep_latest prune strategy for ThreadTTLConfig and checkpointer config passthrough.

└──▷ GET THIS VERSION
$ git clone --branch cli==0.4.14 https://github.com/langchain-ai/langgraph.git
# already have the repo? check out this version:
$ git checkout cli==0.4.14
  • Adds keep_latest prune strategy to ThreadTTLConfig, giving finer control over which threads are retained when TTL cleanup runs.
  • Passes checkpointer config through to the CLI, enabling checkpointer settings to be applied via CLI invocation.
Was this useful?

OpenAI Agents SDK

Sources Release notes → v0.10.3 NOTES

OpenAI Agents SDK v0.10.3 exposes agent tool invocation metadata and a new tool_context accessor on RunResult.

└──▷ GET THIS VERSION
$ git clone --branch v0.10.3 https://github.com/openai/openai-agents-python.git
# already have the repo? check out this version:
$ git checkout v0.10.3
  • Adds tool_context accessor on RunResult to retrieve agent tool invocation context during a run.
  • Exposes immutable agent tool invocation metadata on run results, making per-tool call details available after execution.
Was this useful?

OpenClaw

Sources Release notes → v2026.3.1 NOTES

Your own personal AI assistant. Any OS.

OpenClaw v2026.3.1 adds Android node commands, Discord/Telegram session controls, OpenAI WebSocket-first streaming, Feishu tooling, and a new diff-rendering plugin.

└──▷ GET THIS VERSION
$ git clone --branch v2026.3.1 https://github.com/openclaw/openclaw.git
# already have the repo? check out this version:
$ git checkout v2026.3.1
└──▷ TRY IT
Find exactly which config file OpenClaw is reading — useful when debugging unexpected settings across multiple environments.
$ openclaw config file
Enable the Feishu chat tooling so agents can query chat info and membership within a Feishu channel.
yaml
channels:
  feishu:
    tools:
      chat: true
  • Adds openclaw config file subcommand to print the active config file path resolved from OPENCLAW_CONFIG_PATH or the default location.
  • Adds built-in HTTP liveness/readiness endpoints (/health, /healthz, /ready, /readyz) for Docker/Kubernetes container health checks with fallback routing.
  • Adds Android node commands: camera.list, device.permissions, device.health, notifications.actions (open/dismiss/reply), system.notify, photos.latest, contacts.search/contacts.add, calendar.events/calendar.add, and motion.activity/motion.pedometer.
  • Replaces Discord thread fixed-TTL lifecycle with inactivity (idleHours, default 24 h) and optional hard maxAgeHours controls; adds /session idle and /session max-age commands.
  • Adds per-DM Telegram topic config (direct + topic allowlists, dmPolicy, skills, systemPrompt, requireTopic) with topic-aware authorization and debounce.
+14 moreshow less
  • Makes OpenAI Responses transport WebSocket-first by default (transport: "auto" with SSE fallback) and adds optional WS warm-up via params.openaiWsWarmup.
  • Adds new optional diffs plugin tool for read-only diff rendering from before/after text or unified patches, with gateway viewer URLs for canvas and PNG output.
  • Adds Feishu Docx table actions (create_table, write_table_cells, create_table_with_values) and file/image upload actions (upload_image, upload_file).
  • Adds Feishu inbound reaction handling (im.message.reaction.created_v1) with verified routing through synthetic inbound turns.
  • Adds feishu_chat tool actions for chat info and member queries, configurable under channels.feishu.tools.chat.
  • Supports optional owner permission grant fields on feishu_doc create.
  • Adds German (de) locale support and auto-renders language options from supported locale constants in the Web UI Overview settings.
  • Supports custom OpenAI baseUrl and embedding dimensions for LanceDB memory.
  • Adds opt-in lightweight bootstrap mode for cron/heartbeat automation runs via --light-context flag and agents.*.heartbeat.lightContext config key.
  • Sets OPENCLAW_SHELL environment variable across shell-like runtimes (exec, acp, acp-client, tui-local) so shell startup rules can target OpenClaw contexts.
  • Replaces ad-hoc subagent completion handoff with typed internal task_completion events rendered consistently across direct and queued announce paths.
  • Sets adaptive as the default thinking level for Anthropic Claude 4.6 models (including Bedrock Claude 4.6 refs).
  • Localizes cron page labels, filters, form help text, and validation/error messaging in English and zh-CN.
  • Adds ACP final_only stream delivery default and configurable ACPX command/version probing, pinned to plugin version 0.1.15.
Was this useful?

PydanticAI

Sources Release notes → v1.64.0 NOTES

PydanticAI v1.64.0 adds template=False on PromptedOutput and NativeOutput to suppress schema prompts.

└──▷ GET THIS VERSION
$ git clone --branch v1.64.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v1.64.0
└──▷ USE IT
Suppress the auto-generated schema prompt on a structured output when you are supplying your own formatting instructions.
python
from pydantic_ai import PromptedOutput

output = PromptedOutput(MyModel, template=False)
  • Adds template=False parameter to PromptedOutput and NativeOutput to disable automatic schema prompt injection when you want full control over the model prompt.
Was this useful?

Microsoft Semantic Kernel

Sources Release notes → python-1.40.0 NOTES

Semantic Kernel Python 1.40.0 adds support for (Azure) OpenAI realtime audio models.

└──▷ GET THIS VERSION
$ git clone --branch python-1.40.0 https://github.com/microsoft/semantic-kernel.git
# already have the repo? check out this version:
$ git checkout python-1.40.0
  • Adds support for OpenAI and Azure OpenAI realtime audio models in the Python SDK.
Was this useful?
◆  AI Model & Data Infrastructure

Ollama

Sources Release notes → v0.17.5 NOTES

Get up and running with Kimi-K2.6, GLM-5.2, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Ollama v0.17.5 adds Qwen3.5 model support and peak memory visibility in the MLX engine.

└──▷ GET THIS VERSION
$ git clone --branch v0.17.5 https://github.com/ollama/ollama.git
# already have the repo? check out this version:
$ git checkout v0.17.5
└──▷ TRY IT
Pull and run a small Qwen3.5 model while monitoring peak MLX memory usage during inference.
$ ollama pull qwen3.5:4b && ollama run --verbose qwen3.5:4b
  • Adds Qwen3.5 model series (0.8B, 2B, 4B, 9B parameters) to the Ollama library.
  • ollama run --verbose now shows peak memory usage when using Ollama's MLX engine.
Was this useful?
◆  AI Coding Agents

OpenAI Codex CLI

Sources Release notes → rust-v0.107.0 NOTES

Lightweight coding agent that runs in your terminal

Codex CLI gains thread forking for sub-agents, configurable memories, multimodal custom tools, and richer voice device controls.

└──▷ GET THIS VERSION
$ git clone --branch rust-v0.107.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.107.0
└──▷ TRY IT
Hard-wipe all saved memory state when you want a clean slate — useful after major project switches or corrupted memory accumulation.
$ codex debug clear-memories
  • Adds thread forking into sub-agents, letting you branch work from the current conversation without losing context.
  • Enables microphone and speaker device selection for realtime voice sessions, with persistent device preferences and improved transcription-aligned audio format.
  • Supports multimodal output (including images and structured content) from custom tools, beyond plain text.
  • Introduces configurable memories and a new codex debug clear-memories command to fully reset saved memory state.
  • Exposes richer model availability and plan-upgrade metadata in the app server, surfacing plan-gated models with tooltips in the TUI.
+1 moreshow less
  • Adds an ephemeral field to the Thread object in the app-server API.
Was this useful?

Earendil Works Pi

Sources Release notes → v0.55.4 NOTES

AI agent toolkit: unified LLM API, agent loop, TUI, coding agent CLI

Pi v0.55.4 adds live tool registration, per-tool system-prompt injection via promptSnippet and promptGuidelines.

└──▷ GET THIS VERSION
$ git clone --branch v0.55.4 https://github.com/earendil-works/pi.git
# already have the repo? check out this version:
$ git checkout v0.55.4
└──▷ USE IT
Dynamically register a tool mid-session so the LLM can use it immediately without restarting or running /reload.
typescript
pi.registerTool({
  name: "lookup",
  description: "Look up a value",
  promptSnippet: "lookup: fetches live data by key",
  promptGuidelines: "Use lookup when the user asks for real-time data.",
  execute: async (args) => { /* … */ }
});
  • Enables runtime tool registration that takes effect immediately in active sessions — tools added via pi.registerTool() are available to pi.getAllTools() and the LLM without /reload.
  • Adds promptSnippet field to ToolDefinition for injecting a one-line entry into the default system prompt's Available tools section while the tool is active.
  • Adds promptGuidelines field to ToolDefinition so active tools can append tool-specific bullets to the default system prompt's Guidelines section.
  • Supports custom tool renderers that suppress transcript output cleanly, leaving no blank rows or empty footprint in interactive rendering.
Was this useful?
Other / Uncategorized
◆  AI OBSERVABILITY

Arize Phoenix

Sources Release notes → arize-phoenix-v13.7.0 4 RELEASES · 2026-03-02 NOTES STABLE

Arize Phoenix 13.7.0 updates the container image base to Debian 13.

└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v13.7.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v13.7.0
  • Updates the official container image base to Debian 13.
3 more releases in this issue · 2026-03-02
arize-phoenix-otel-v0.15.0 NOTES STABLE

arize-phoenix-otel now supports Python 3.10 through 3.14

└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-otel-v0.15.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-otel-v0.15.0
  • Extends arize-phoenix-otel compatibility to Python 3.10, 3.11, 3.12, 3.13, and 3.14.
arize-phoenix-client-v1.30.0 NOTES STABLE

Phoenix client v1.30.0 adds GET endpoints for the sessions REST API.

└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-client-v1.30.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-client-v1.30.0
  • Adds GET endpoints for the sessions REST API, enabling programmatic retrieval of session data.
arize-phoenix-v13.6.0 NOTES STABLE

Phoenix v13.6.0 adds inline document annotation for retriever spans and CLI sessions commands for multi-turn conversations.

└──▷ GET THIS VERSION
$ git clone --branch arize-phoenix-v13.6.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v13.6.0
  • Adds sessions and session subcommands to the CLI for multi-turn conversation support.
  • Adds inline document annotation UI for retriever spans in the app.
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 →