How MultimodelChat works

Create an API key in your provider's dashboard, then add it in Settings → AI Providers. Multimodel Chat tests the key live against the provider before saving it, so a typo or billing problem surfaces immediately instead of failing mid-conversation. Each guide below walks through account creation, billing, and the exact key page.

Once a provider is connected, every model it exposes appears in the model selector — GPT-5.6 Sol and GPT-5.4 Mini, Claude Opus 5 and Sonnet 5, Gemini 3.1 Pro and 3.7 Flash, Grok 4.6, and 500+ models through a single OpenRouter key. Set a default model per conversation, or pick a different one per message. Compare everything in features.

Draft with GPT, ask Claude to review it, then have Grok research the latest information — all in one thread. Multimodel Chat sends the conversation's supported message history to the newly selected provider, so text context carries over without copy-pasting. Provider- specific reasoning, tool state, unsupported attachments, and content beyond a model's context window may not transfer identically. For when switching helps (and when it costs more), read how to switch between AI models mid-conversation.

Your key stays in this browser. It is encrypted at rest with AES-256 in the browser's vault, sent over TLS with each request, and used in memory — it is not stored on our servers. Browser storage is tied to one browser profile and does not protect against an active browser compromise.