CORS Chat is a browser-based tool designed to help test OpenAI-Responses-compatible chat endpoints. The developer built it today using GPT-5.6-Sol xhigh and used it to exercise Qwen 3.8 27B running in LM Studio on both an M5 MacBook Pro and an NVIDIA DGX Spark.
What the tool does
- Provides a web UI for invoking and testing chat endpoints that follow the OpenAI-Responses compatibility.
- Persists conversations in the browser and allows export as copy-pasteable JSON.
- Detects generated SVG images and progressively renders them in the chat while tokens are still streaming.
Tested environments
The developer reports testing the tool against:
- Qwen 3.8 27B in LM Studio on an M5 MacBook Pro and on an NVIDIA DGX Spark.
- LM Studio started with the --cors option, and OpenRouter endpoints; both setups worked.
Why it matters
CORS Chat provides a lightweight, browser-based way to validate and interact with LLM chat endpoints without needing a dedicated client application. Local persistence and JSON export help with reproducibility and debugging. The progressive SVG rendering is useful when models emit visual output incrementally, because it shows real-time progress while tokens stream.
Tags
svg, ai, generative-ai, llms, cors, openrouter, lm-studio
Note on creation
The developer states the tool was built today with GPT-5.6-Sol xhigh; the compatibility and behavior described reflect the tests performed during that development session.
