AI models, search providers, and API keys
1. Search works out of the box with free providers (DuckDuckGo, Wikipedia, Semantic Scholar, arXiv).
2. Add your own API keys in Search APIs tab to unlock premium providers (Tavily, Exa, Firecrawl, Brave).
3. Add local or cloud models for AI-powered planning, validation, and synthesis.
Models are optional. Free search providers are always active.
Auto-detect models running on your machine. Use this for Ollama and LM Studio. Toggle multiple models to enable parallel reasoning.
Start Ollama and pull a model to use it here. Run: ollama pull llama3.2
Start LM Studio and load a model to use it here. Make sure the LM Studio local server is started and answering on /v1/models.