Smalt Docs

Get your keys

With Smalt thinking, it uses one cloud model with your own API key: Claude, GPT, Mistral, or any OpenAI-compatible server. A TypeSafe key is optional.

Why your own key?

All of it is in Settings → Cloud → Cloud model and keys: pick the Provider, then paste its key.

Anthropic (Claude)

  1. Go to console.anthropic.com and sign in.
  2. Add a little credit under Billing: API use is paid separately from a Claude subscription.
  3. Open API keys → Create Key, name it "Smalt" and copy it. It starts with sk-ant-.
  4. In Smalt: Provider Anthropic, then paste the key.

OpenAI (GPT)

  1. Go to platform.openai.com/api-keys and sign in.
  2. Add credit under Billing: API use is paid separately from a ChatGPT subscription.
  3. Click Create new secret key, name it "Smalt" and copy it. It starts with sk-.
  4. In Smalt: Provider OpenAI, then paste the key.

Mistral

  1. Go to console.mistral.ai and sign in.
  2. Set up billing for your workspace.
  3. Open API Keys, create a key, name it "Smalt" and copy it.
  4. In Smalt: Provider Mistral, then paste the key.

Other (OpenAI-compatible)

For a server that speaks the OpenAI API, on your Mac or elsewhere: Provider Other (OpenAI-compatible), then its Server URL and the model name.

http://localhost:11434/v1

The key is optional for a server on your Mac.

TypeSafe (optional, spots questions)

  1. Go to console.typesafe.ai/keys and sign in.
  2. Create a key and copy it.
  3. In Smalt: Settings → Cloud → Live answers → Jev, then paste it into TypeSafe key.

Jev judges every line of the meeting in about a third of a second, for a few cents a meeting. Without it, the cloud model does.

Check it

Click Check next to the key. It says Works, or what's wrong. Settings → Voice Assistant shows the same under Check Keys.