Works with / Mistral

Synapse with Mistral

Run your Bots on Mistral models with your own Mistral API key.

Have a Mistral Vibe plan instead? See your coding subscription.

What works

• Models: Mistral Medium 3.5, Mistral Large 3 and Mistral Small 4. Synapse follows Mistral's latest version of each.
• Chat, tools and replies, through the same approvals as every Bot.
• Images on Medium and Large. Small doesn't take images.
• Voice calls.
• Subagents, on the same model.
• Computer and browser helpers, on the same model. Medium and Large see the screen in screenshots; Small works from text reads of the screen and page instead, as Experimental.
• Coding agents and Engineering mode, on the same model, through the same approvals.

Every model starts as Not checked. A badge is earned by measurement, not given. See how badges are earned.

Not yet

• Web search of its own. A Mistral Bot borrows search from OpenAI, Gemini or OpenRouter if you've set one up; otherwise it has none.

Cost

Mistral bills your own key. Synapse meters every call and counts it against your spend budget. When the budget is reached, the model isn't called.

US dollars per million tokens. Mistral's list prices, checked on Sep 30, 2026.

• Mistral publishes no separate cached-input price, so Synapse counts cached input at the full input price.
• Mistral doesn't report usage on a streamed reply, so Synapse estimates it from the length of the text.
• Mistral's pages don't state the context windows. Synapse assumes 128,000 tokens.
• Synapse's own small tasks, like titles and key checks, use Mistral Small 4.
• The model picker shows what 100 turns like your Bot's recent ones would cost on each model.

Privacy

• Consent first. Nothing goes to Mistral until you allow it once. The consent says what is sent: the conversation, tool results, files the Bot reads, safety-check inputs, memory notes and call transcripts. Mistral's own terms apply.
• Your key stays on your Mac. It is encrypted, and only Synapse's own proxy adds it to a request. Your Bots never see it. Each call gets a one-time token, and the key is removed from anything that comes back.
• The same safety gate. Every action goes through Synapse's fixed rules and approvals, whatever the model. The safety reviewer runs on Claude when you've saved an Anthropic key, and otherwise on the provider you set up. You can pick another model for it in Settings → Rules → Safety reviewer. A reviewer on a model other than Claude asks you first until it passes the safety check.

Set up

1. Create an API key in the Mistral console.
2. In Synapse, open Settings → Account → Model providers. Click Allow next to Mistral, read what it sends, and allow it.
3. Paste the key, click Save, then Test key.
4. Open a Bot's settings and pick a Mistral model.

You can also start with Mistral: pick it in the key step when you first open Synapse, allow it, and paste the key. It's tested before it's saved, and no Anthropic key is needed.