Connect local models first, or add a cloud API.
Run a model on your own computer with Ollama - free, private, and works offline.
ollama pull qwen2.5:3b
./start-ollama.sh
Rather type it yourself? Same thing in one line - OLLAMA_ORIGINS is what lets the browser reach it:
pkill -f ollama; OLLAMA_ORIGINS=* ollama serve
http://127.0.0.1:11434/v1
Use the numbers, not localhost: on Windows localhost usually resolves to IPv6 ::1 first, while Ollama listens on IPv4 only - so localhost gets refused before it ever reaches the server.
Tip: bigger models (llama3.1:8b, qwen2.5:7b) are smarter but need more RAM. Start small and scale up.
Manage the endpoints you've added.
We can't reach Ollama on this computer yet.
One-click downloads pull the model straight through Ollama, so it needs to be installed and running first.
Already have it? Run the bundled script from the project folder - ./start-ollama.sh - it frees the port and starts Ollama with browser access in one go, then retry below. Prefer one line? pkill -f ollama; OLLAMA_ORIGINS=* ollama serve.