The Wire · Showcase
CODEX CLEANS UP ASYNC CHAOS WHILE SDK TWINS SHIP TRANSCRIPTION UPDATES
By RepoJournal · Filed · About OpenAI
Codex ditches the wait for side conversation cleanup, routing it to the background while your UI stays responsive, as both Node and Python SDKs land transcription model updates in lockstep.
The Codex team shipped a critical responsiveness fix that removes side conversations from local TUI state immediately when switching threads, then handles interrupt and unsubscribe cleanup in the background [1]. This pairs with deeper infrastructure work tracking parent turns through nested Codex requests and agent spawns, preserving provenance across delegated sessions and queued messages [2]. Meanwhile, the V8 runtime got refreshed to rusty_v8 150.4.0, updating Bazel targets and LLVM source revisions for the new release [3]. On the SDK front, openai-node v7.1.0 and openai-python v2.50.0 shipped simultaneously with transcription model updates [6] [7], though Python had to restore a transcription keyword overload that got lost in the process [7]. Codex also standardized its HTTP client usage across TUI checks and MCP OAuth flows, ditching ad-hoc reqwest calls in favor of the shared configured client pool [4] [5].
One email a day. Unsubscribe in one click.
Keep up with OpenAI in about 3 minutes a day: what actually shipped — the commits, pull requests, releases, and security advisories that matter.
One email a day. Unsubscribe in one click. Read a past issue →
Action items
- → Update openai-node to v7.1.0 and openai-python to v2.50.0 for transcription model support openai/openai-node [plan]
- → Deploy Codex changes for responsive side conversation cleanup on next release openai/codex [plan]
- → Monitor Codex parent turn tracking in nested request scenarios openai/codex [monitor]
References
- [1] Clean up side conversations in the background ↗ openai/codex
- [2] Track parent turns for nested Codex requests ↗ openai/codex
- [3] Update rusty_v8 to 150.4.0 ↗ openai/codex
- [4] Use the shared HTTP client for TUI network checks ↗ openai/codex
- [5] Use configured HTTP clients for all MCP OAuth requests ↗ openai/codex
- [6] v7.1.0 ↗ openai/openai-node
- [7] v2.50.0 ↗ openai/openai-python