Skip to content

Turnspan 6

Choose a tag to compare

@michaelgiraldo michaelgiraldo released this 21 Nov 06:56
0d916c0

Package openai-chatkit 1.6.5+turnspan.6 · Python 3.14 · Tag v1.6.5+turnspan.6

Source commit 0d916c03ef8e48d4d77ddcc461431e1d6f7b8718.

  • Breaking store contract. Every streaming response has a durable turn
    ID. Custom stores implement atomic turn ownership, fenced responder
    writes, and persisted milestones. Use a fresh schema for this candidate.
    Streaming actions receive turn_id. Supply the current ID to
    update_thread and other direct responder writes.
  • The server distinguishes a saved callback from responder start and
    completion. Closing a response before processing starts records
    ABANDONED. Started responders can report SUPERSEDED when a new
    request takes ownership.
  • Breaking Python and storage contract. Browser tools publish a
    ClientToolCallItem containing display data. The server registers a
    separate ClientToolInvocation with model data and the turn linkage.
    Custom responders yield ClientToolRequest to register the pair.
    Pass the complete ClientToolTurn to
    client_tool_continuation_input. The ChatKit browser still sends the
    same callback envelope.
  • Browser tools return the ID of the exact saved invocation. Each callback can
    complete once, and a newer turn can cancel a pending invocation.
  • The Agents SDK adapter preserves the original model call and any sibling tool
    results. Choose local replay or OpenAI-managed Responses context when
    registering the tool.
  • The browser tool guide
    documents the current callback, continuation, and storage contracts.

Install

python -m pip install 'openai-chatkit[agents,fastapi] @ https://github.com/michaelgiraldo/chatkit-python/archive/0d916c03ef8e48d4d77ddcc461431e1d6f7b8718.tar.gz'

Documentation for this commit · Pin and verify the installed build

Verification

  • Passing CI run

  • Locked versions in the SDK repository were OpenAI Python 3.56.0 and Agents SDK 0.25.0.

  • The consumer smoke ran on Python 3.14.10 with OpenAI Python 3.57.0 and Agents SDK 0.25.0. It exercised the quickstart, FastAPI, and mocked model HTTP from an isolated source install.

  • GitHub release verification installed the public source archive and passed all 30 smoke tests.