Ollama: Codex Integration Push and Schema Hardening
This cycle centered on deepening Codex and Qwen Code integration inside the Ollama app, alongside a cluster of fixes tightening how tool schemas and model options are handled internally. Several bugs around context length, escaping, and shared state got closed out with clear root-cause explanations.
Duration: PT2M58S
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-09-05T13:01:18Z
- Audio duration: PT2M58S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning. It's September 5th, and today's codebase activity tells a clear story: Ollama is investing heavily in becoming the backend of choice for third-party coding agents.
The biggest theme is Codex and Qwen Code alignment. Parth Sareen shipped a run of related changes: routing Codex's auto review through the selected model in PR 18246, hardening the Codex desktop proxy in PR 18244, and preserving Codex configuration across host changes in PR 18247. On top of that, PR 18224 adds…
The second theme is correctness in what gets forwarded to the model. PR 18248 fixes a subtle bug where escaped pattern characters in nested tool schema fields survived encoding while top-level ones were silently dropped — a real behavioral asymmetry between API surfaces. PR 18237 fixes HTML characters getting…
A smaller but important fix: PR 18253 deep-clones cached model options so nested arrays and maps in one request can't mutate shared inference cache state — a concurrency correctness fix confirmed with a race-detector test run.
Rounding out the cycle: Daniel Hiltgen bumped MLX and bounded the macOS MLX build in CI, PR 18243 blocks oversized model pulls by default with a force override, and PR…
What…
Nearby episodes from Ollama
- Structured Output Meets Speculative Decoding
- Hardening the MLX Engine and Fixing What Feeds the Scheduler
- MLX Grows Up, Model Intent Gets Respected
- Runner Reliability and a Documentation Cleanup Wave
- Proxy Fix and a Documentation Cleanup Sweep
- Weekly Recap - Claude Desktop Integration & MLX Engine Maturity
- Correctness Fixes Under the Hood
- Agent Expansion, Desktop Polish, and a Documentation Sweep