All pull requests

Pull requests

• feat(rime): add Coda custom_lexicon option

  #8245·saad-py opened Oct 29, 2026·

  ·
  1

  5

• fix: wait for in-flight acquisitions when closing connection pools

  #8244·maximilliangrand opened Oct 29, 2026·

  ·

• fix(amd): settle uncertain when the callee hangs up while listening

  #8242·sarankumar1325 opened Oct 29, 2026Contributor·

  ·
  1

  3

• fix(llm): select container union defaults by item schema

  #8240·Ramcharan49 opened Oct 29, 2026·

  ·

• fix(aws): pass chat(extra_kwargs=...) to the Bedrock converse_stream request

  #8239·pcbeingused333 opened Oct 29, 2026Contributor·

  ·

• fix(blingfire): build CPython 3.15 wheels from public source

  #8238·GriffinCanCode opened Oct 29, 2026·

  ·
  5

• fix(llm): give a raw tool with empty parameters an object schema for Anthropic and Bedrock

  #8235·jaeminbird opened Oct 29, 2026·

  ·
  1

• fix(llm): send a developer-role preamble as the system prompt to Anthropic, Google and AWS

  #8234·jaeminbird opened Oct 29, 2026·

  ·
  1

• fix(openai): yield to the event loop after each realtime client event

  #8228·zek opened Oct 28, 2026·

  ·
  1

  1

• fix(telemetry): keep the stall sampler out of linecache

  #8226·zek opened Oct 28, 2026·

  ·
  1

  8

• fix(tts): free a dropped TTS without the cyclic garbage collector

  #8224·zek opened Oct 28, 2026·

  ·
  1

  5

• fix(aio): free a closed Chan without the cyclic garbage collector

  #8222·zek opened Oct 28, 2026·

  ·
  1

  1

• fix(voice): preserve task results after parent speech

  review_effort:high
  #8220·chenghao-mou opened Oct 28, 2026Member·

  ·
  3

• fix(openai): end a failed realtime response without an unretrieved exception

  #8218·zek opened Oct 28, 2026·

  ·
  1

  2

• feat(examples): add the four homepage demo agents

  #8216·ShayneP opened Oct 28, 2026Contributor·

  ·
  14

• docs(inference): Example for inference.responses.LLM

  inference-review
  #8213·adrian-cowham opened Oct 28, 2026Contributor·6/6·

  ·
  6

  

• feat(inference): update_options and response_format for inference.responses.LLM

  inference-review
  #8212·adrian-cowham opened Oct 28, 2026Contributor·5/6·

  ·
  18

  

• feat(inference): WebSocket transport for inference.responses.LLM

  inference-review
  #8211·adrian-cowham opened Oct 28, 2026Contributor·4/6·

  ·
  36

  

• feat(inference): Add inference.responses.LLM

  inference-review
  #8210·adrian-cowham opened Oct 28, 2026Contributor·3/6·

  ·
  5

  

• fix(openai): Keep only the first message of a Responses reply

  inference-review
  #8209·adrian-cowham opened Oct 28, 2026Contributor·2/6·

  ·
  7

  

• feat(gnani): add a model option to STT (Prisma v3.0)

  #8207·adsk2050 opened Oct 28, 2026·

  ·
  1

• a2a: the live inbox, a busy agent in text

  #8204·longcw opened Oct 28, 2026Contributor·6/6·

  ·
  5

• a2a: an answer continues the task that asked

  #8203·longcw opened Oct 28, 2026Contributor·5/6·

  ·
  4

• a2a, delegation: end the task when a turn, a worker or an expert fails

  #8202·longcw opened Oct 28, 2026Contributor·4/6·

  ·
  4

• delegation: send exact text when it is said, and read the answer from the artifact

  #8201·longcw opened Oct 28, 2026Contributor·3/6·

  ·
  5