Connect another AI provider

The flow is the same for every provider — in the @bridge Settings → AI providers tab, pick the provider, paste your API key, and click Connect. Your key is encrypted on this device and never reaches atbridge.ai’s servers. The only thing that differs is where you get the key.

Note

Install atbridge.ai and the Thunderbird add-on first — see Install atbridge.ai. New to the in-Thunderbird side? Start with Set up the @bridge extension.

Where to get each key

Pick your provider below, create a key on its site, then paste it into the AI providers tab.

• Mistral AI — create a key in the Mistral admin console.

• Groq — fast Llama (and others); create a key in the Groq console.

• DeepSeek — create a key in the DeepSeek platform.

• Kimi (Moonshot AI) — the powerful Kimi K3 models; create a key in the Moonshot platform.

• xAI (Grok) — open the xAI console, then API Keys in the sidebar (the direct link is tied to your team). Your account needs credits.

• OpenRouter — one key, many models; create a key on OpenRouter.

• NVIDIA NIM — free to try, no credit card: create a key at build.nvidia.com (open any model → Get API Key). A great way to try @bridge AI Chat with no provider bill (your atbridge.ai plan still applies) — but mind the trade-offs below.

• OpenAI-compatible — any endpoint that speaks the OpenAI API (a self-hosted or proxy endpoint). Pick OpenAI-compatible in the Provider dropdown, then fill the Base URL field in addition to the API key.

NVIDIA NIM: free, but NOT private and rate-limited

NVIDIA’s free tier is a great way to try @bridge AI Chat without a provider bill (your atbridge.ai plan still applies) — but understand what you’re trading:

• Not private. Your prompts run on NVIDIA’s cloud, not your machine — unlike a local Ollama model. Don’t use it for content you need to keep off third-party servers.
• Rate-limited. The free tier caps requests (around 40 requests per minute), so heavy or bursty use can hit a short cooldown.

For free and private, run a local model with Ollama instead.

Caution

Each paid provider bills your own account per use — atbridge.ai puts no markup on what the provider charges. Your keys stay encrypted on this device. Prefer not to pay per use? Run a model free and locally with Ollama.

Add it to atbridge.ai

1. In Thunderbird, open the @bridge Settings — the gear, or Tools → Add-ons and Themes → atbridge.ai → Options (see Set up the @bridge extension). Go to the AI providers tab.

2. Pick your provider from the Provider dropdown. For OpenAI-compatible, also fill in the Base URL field.

3. Paste your API key into API key. It’s encrypted on this device — your key never reaches atbridge.ai’s servers.

4. Click Connect. The status turns to a green ✓.

Pick a model and chat

Open the @bridge AI Chat (the @b button in Thunderbird) and pick a model from the model dropdown — for most providers atbridge.ai fetches the live model list, so you choose from what your account can actually use — then ask away:

“Summarise the unread thread from Maya and draft a reply.”

Tip

You can connect more than one provider and switch in the chat any time — your keys all stay local.

Related

• Connect your AI — overview
• Connect OpenAI
• Use the @bridge AI Chat