Simon Willison released CORS Chat, a web tool for testing local LLMs
Simon Willison created the CORS Chat web tool for testing local LLMs (e.g. Qwen 3.8 27B in LM Studio) as well as models via OpenRouter, with conversations saved in the browser and progressive rendering of generated SVG images.
Independent commentator Simon Willison created a web tool called CORS Chat that serves as a user interface for OpenAI-Responses-compatible chat endpoints. According to him, the tool was built in one day with the help of the GPT-5.6-Sol model in xhigh reasoning mode, specifically to test the Qwen 3.8 27B model running in LM Studio on two different machines – a MacBook Pro with an M5 chip and an NVIDIA DGX Spark device.
Willison tested CORS Chat against both LM Studio running with the --cors option and the OpenRouter service, and according to him, it worked without any problems in both cases. Conversations are saved directly in the browser and can be exported as JSON for later processing or sharing.
The tool also includes a feature that detects when a model is generating an image in SVG format and renders that image progressively while individual tokens are still arriving from the stream.
Why it matters
The tool removes the need for a custom UI for anyone who wants to quickly try a model running locally in LM Studio or a model available via OpenRouter – all that is needed is an OpenAI-compatible endpoint. This has practical uses for developers and enthusiasts experimenting with local LLM deployment on their own hardware, including watching streamed SVG image generation in real time.
Relevant practical impact
What this means
For individuals
Anyone running LLMs locally via LM Studio or using OpenRouter gets a simple browser-based chat tool with support for streamed SVG and conversation exports to JSON, without having to build their own UI.
Check the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.