Who contributed to OpenAI on May 4, 2026?

2 developers shipped this update, including pakrym-oai and etraut-openai.

What were the notable OpenAI updates?

[codex] Emit MCP tool calls as turn items, [codex] Refactor app-server dispatch result flow, and Remove local docs and specs.

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

CODEX CONSOLIDATES TOOL CALL OWNERSHIP INTO CORE TURNITEM

By RepoJournal · Filed 06:02 UTC on May 4, 2026 · About OpenAI

2 people shipped this

pakrym-oai @pakrym-oai 2 cited

etraut-openai @etraut-openai 1 cited

McpToolCall migrations are complete — app-server now emits tool calls as canonical turn items instead of synthesizing them from deprecated legacy events, collapsing three separate changes into a unified dispatch and response model.

MCP tool calls moved into core `TurnItem`s [1], closing the gap between app-server synthesis and v2 `ThreadItem::McpToolCall` shape. This keeps MCP result semantics in one place instead of spread across conversion code, and gives core ownership over the lifecycle — it now decides whether a completed call is `completed` or `failed`. Simultaneously, app-server dispatch refactored response sending [2]: normal `ClientResponsePayload` sends centralized in the dispatch path, while explicit-response methods stay explicit where they need custom ordering or delayed delivery. This unwinds handler wrappers that were making it impossible to see which requests return payloads and which emit notifications after response. On the documentation front [3], local-only docs and planning specs are being removed from the repository — `docs/` now links to the canonical surface on developers.openai.com instead of duplicating transient implementation notes.

Action items

→ Review dispatch flow changes in your MCP integrations — response ordering may have shifted openai/codex [plan]
→ Update local tooling pointing to docs/ — redirect to developers.openai.com openai/codex [monitor]

References

[1] [codex] Emit MCP tool calls as turn items ↗ openai/codex
[2] [codex] Refactor app-server dispatch result flow ↗ openai/codex
[3] Remove local docs and specs ↗ openai/codex

Quick answers

What shipped in OpenAI on May 4, 2026?: McpToolCall migrations are complete — app-server now emits tool calls as canonical turn items instead of synthesizing them from deprecated legacy events, collapsing three separate changes into a unified dispatch and response model. In total, 4 commits and 4 pull requests landed.
Who contributed to OpenAI on May 4, 2026?: 2 developers shipped this update, including pakrym-oai and etraut-openai.
What were the notable OpenAI updates?: [codex] Emit MCP tool calls as turn items, [codex] Refactor app-server dispatch result flow, and Remove local docs and specs.

CODEX HARDENS MCP RUNTIME STABILITY ACROSS FIVE CRITICAL FIXES

OpenAI shipped a coordinated hardening of the Model Context Protocol layer across Codex, patching race conditions in thread startup, resource access, and authority management that could leave stale configurations in production.

ai 66 shipped 1-min read

OpenAI 1 day ago

SDK RELEASES LOCK IN PROMPT CACHING AND SPEND CONTROLS

Both Node and Python clients shipped v6.49.0 and v2.48.0 overnight with admin spend limit APIs and relaxed prompt cache validation.

ai 103 shipped 1-min read

OpenAI 2 days ago

CODEX CONSOLIDATES THREAD STARTUP, NODE SDK ADDS REALTIME AND STANDARD SCHEMA SUPPORT

Codex refactored its threading and database layer for consistency while the Node SDK ships four new schema helpers and fixes audio streaming in a single release push.

ai 120 shipped 1-min read

OpenAI 3 days ago

CODEX FIXES WINDOWS EXECUTION BUGS WHILE TIGHTENING SECURITY REVIEW GATES

Codex shipped critical fixes for Windows process management and terminal navigation that were silently breaking execution sessions on Windows machines.

ai 104 shipped 1-min read

Elsewhere on the wire

AI Agents about 9 hours ago

CLAUDE OPUS 5 LANDS ACROSS THE STACK

The newest Anthropic model is now live in langchain, Cline, and llama-index, with native support for extended reasoning and 1M context windows.

ai-agents 28 shipped 1-min read

Local LLMs about 9 hours ago

OLLAMA LANDS LAGUNA SUPPORT AND CRUSHES MEMORY LEAKS WHILE SGLANG HITS V0.5.16 WITH CONFIDENCE-DRIVEN SPECULATIVE DECODING

Ollama shipped three critical performance and reliability fixes for Metal residency and concurrent access patterns, while SGL-Lang released 0.5.16 with a new speculative algorithm hitting 383.7 tok/s on DeepSeek-V4.