Backfill issue -355 — 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 Agent Frameworks

LangChain

langchain-core==0.3.75 NOTES

LangChain Core 0.3.75 adds response body tracing on errors for easier debugging.

└──▷ GET THIS VERSION

$ git clone --branch langchain-core==0.3.75 https://github.com/langchain-ai/langchain.git
# already have the repo? check out this version:
$ git checkout langchain-core==0.3.75

›Traces the response body when an error occurs, giving practitioners visibility into what the model returned at the point of failure.

Was this useful?

Letta (formerly MemGPT)

0.11.5 NOTES

Letta 0.11.5 adds background mode for message streaming.

└──▷ GET THIS VERSION

$ git clone --branch 0.11.5 https://github.com/letta-ai/letta.git
# already have the repo? check out this version:
$ git checkout 0.11.5

›Adds background mode for message streaming, enabling stream consumption without blocking the caller.

Was this useful?

PydanticAI

v0.8.0 2 RELEASES · 2025-08-26 NOTES STABLE

PydanticAI v0.8.0 adds elicitation callbacks for MCP servers, message history in CLI agents, and a richer AgentStreamEvent union type.

└──▷ GET THIS VERSION

$ git clone --branch v0.8.0 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.8.0

└──▷ USE IT

Seed a CLI agent session with prior message history so returning users resume context rather than starting fresh.

python

agent.to_cli(message_history=prior_messages)

›Adds message_history parameter to agent.to_cli() to seed CLI sessions with prior conversation context.
›Adds elicitation callback support to MCP servers, enabling agents to request additional input from users during tool execution.
›Makes AgentStreamEvent a union of ModelResponseStreamEvent and HandleResponseEvent, expanding the event types available when streaming agent responses.

1 more release in this issue · 2025-08-26

v0.7.6 NOTES STABLE

PydanticAI v0.7.6 adds a Cerebras provider and renames the OpenAI model class.

└──▷ GET THIS VERSION

$ git clone --branch v0.7.6 https://github.com/pydantic/pydantic-ai.git
# already have the repo? check out this version:
$ git checkout v0.7.6

›Adds CerebrasProvider (Cerebras provider integration) for running agents against Cerebras-hosted models.
›Replaces all_messages_events with pydantic_ai.all_messages span/event name under InstrumentationSettings(version=2).
›Deprecates OpenAIModel in favor of the new OpenAIChatModel class.

└──▷ BREAKING ON UPGRADE

!The tenacity retry implementation has changed behavior — existing retry logic built on the prior AsyncTenacityTransport semantics may behave differently on upgrade.

Was this useful?

◆ AI Coding Agents

OpenAI Codex CLI

rust-v0.24.0 NOTES

Lightweight coding agent that runs in your terminal

Codex CLI gains web search, transcript mode, image paste, queued messages, and long-running shell command support.

└──▷ GET THIS VERSION

$ git clone --branch rust-v0.24.0 https://github.com/openai/codex.git
# already have the repo? check out this version:
$ git checkout rust-v0.24.0

└──▷ TRY IT

Review the full session transcript with scrolling after a long agentic run to audit every command and output.

$ # Inside a running codex TUI session, press Ctrl+T to open transcript mode and scroll through the history.

Tune GPT-5 response verbosity to reduce noise in automated pipelines.

yaml

# In your codex config file:
model_verbosity = "low"

›Adds web search capability via the explicit --search option.
›New transcript mode (Ctrl+T) with live-updating scrollable view of the full session.
›Supports copy/paste and drag-and-drop of image files directly into the TUI (macOS Terminal + iTerm).
›Introduces queued messages, letting input be staged while the model is still responding.
›Adds edit/resume conversation from any previous message in history (esc-esc).

+7 moreshow less

›Adds long-running shell command support via new exec_command and write_stdin tools.
›New model_verbosity config option for tuning GPT-5 output verbosity via the Responses API.
›Now reads all AGENTS.md files up the directory tree to the git root for richer project context.
›Diff output now displayed in a pager with hunk headers for easier review.
›Simplifies command approval UI in the TUI.
›Hides chain-of-thought (CoT) output by default; surfaces headers in the status indicator instead.
›Improves MCP tool cache hit rate by sorting tools deterministically.

Was this useful?

SST OpenCode

v0.5.28 NOTES

The open source coding agent.

OpenCode v0.5.28 adds a new opencode workflow capability.

└──▷ GET THIS VERSION

$ git clone --branch v0.5.28 https://github.com/sst/opencode.git
# already have the repo? check out this version:
$ git checkout v0.5.28

›Adds opencode workflow support.

Was this useful?

◆ AI Model & Data Infrastructure

NVIDIA Triton Inference Server

v2.60.0 NOTES

Triton Inference Server v2.60.0 adds CUDA 13 support.

└──▷ GET THIS VERSION

$ git clone --branch v2.60.0 https://github.com/triton-inference-server/server.git
# already have the repo? check out this version:
$ git checkout v2.60.0

›Adds CUDA 13 support to Triton Inference Server.

Was this useful?

Other / Uncategorized

◆ AI OBSERVABILITY

Arize Phoenix

arize-phoenix-v11.28.0 2 RELEASES · 2025-08-26 NOTES STABLE

Phoenix v11.28.0 adds a new experiment list page to the frontend UI.

└──▷ GET THIS VERSION

$ git clone --branch arize-phoenix-v11.28.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-v11.28.0

›Adds an experiment list page to the frontend UI for browsing and managing experiments.

1 more release in this issue · 2025-08-26

arize-phoenix-evals-v0.29.0 NOTES STABLE

Phoenix Evals 0.29.0 adds evaluate_dataframe function and evals 2.0 tracing support.

└──▷ GET THIS VERSION

$ git clone --branch arize-phoenix-evals-v0.29.0 https://github.com/Arize-ai/phoenix.git
# already have the repo? check out this version:
$ git checkout arize-phoenix-evals-v0.29.0

└──▷ USE IT

Run LLM evaluations over a pandas DataFrame of traces or records in a single call.

python

from phoenix.evals import evaluate_dataframe

results = evaluate_dataframe(df, evaluators=[my_evaluator])

›Adds evaluate_dataframe function for running evaluations directly against a pandas DataFrame.
›Adds evals 2.0 tracing to instrument and observe evaluation runs.

Was this useful?

◆ MCP TOOLING

Composio

v0.8.9 NOTES

Composio v0.8.9 adds Vercel AI SDK v5 support, a complete CLI login command, and a disable-version-check flag.

└──▷ GET THIS VERSION

$ git clone --branch v0.8.9 https://github.com/ComposioHQ/composio.git
# already have the repo? check out this version:
$ git checkout v0.8.9

›Adds --disable-version-check flag to the CLI to skip version validation on startup.
›Completes the login command in the CLI, enabling full authentication workflows from the terminal.
›Adds support for Vercel AI SDK v5 integration.

Was this useful?

// END OF TRANSMISSION

The AI Toolchain · curated daily · August 26, 2025
Releases & features from the open-source and commercial AI tools that practitioners actually run.

The AI Toolchain — issue -355, August 26, 2025