LLM Chat Completions Server Plugin Enables OpenAI-Compatible Endpoints
On July 30, 2026, Simon Willison announced the launch of llm-chat-completions-server 0.1a0, a plugin for his LLM tool that sets up a localhost server on port 9001. This server provides access to all installed LLM models through an endpoint compatible with OpenAI Chat Completions. Designed to test the new content-addressable logs from LLM 0.32rc1, the plugin aims to eliminate redundancy in conversation states for OpenAI-style requests, where each new message builds on the previous one. The server enables clients to monitor conversation states, with requests progressively increasing in length, utilizing hashes of message components. Notably, the entire plugin was crafted by GPT-5.6 Sol, showcasing its proficiency with the OpenAI Chat Completions API. Willison also introduced a $10/month sponsorship option for a curated email summary of significant LLM updates.
Key facts
- Plugin named llm-chat-completions-server version 0.1a0 released.
- Runs a localhost server on port 9001.
- Exposes full collection of LLM models from installed plugins.
- Uses OpenAI Chat Completions compatible endpoint.
- Built to test content-addressable logs in LLM 0.32rc1.
- Content-addressable logs de-duplicate conversation state using hashes.
- Plugin was written by GPT-5.6 Sol.
- Announced by Simon Willison on 30th July 2026.
Entities
Artists
- Simon Willison
Institutions
- OpenAI