Skip to the page
Conch
DocsGitHub

Providers

Cerebras

Very fast open models

Very fast answers from a small set of open models — trial credit to start, then pay as you go. Conch lends it your integrations.

  • Very fast
  • Open models
  • Uses your integrations
Connects with
A key you paste
Your apps
Conch hands it their tools
A key looks like
csk-…
  • NoYour files and commands
  • YesSaves memories itself
  • YesAsks before each step
  • NoWorks offline

Worth knowing first

  • Tool-capable models can use Conch’s files and connected apps. Commands need the OS sandbox and run without network access; chat-only models cannot take actions.

Connect it

  1. Open Settings → Providers and choose Cerebras.
  2. Press Open cloud.cerebras.ai and create a key. It starts with csk-.
  3. Paste it. Conch checks it before keeping it.

Or paste the key anywhere on Settings → Providers. Conch knows whose key it is from how it starts, and checks it before keeping it.

What you get

Very fast answers from a small set of open models. Trial credit to start, then pay as you go.

Conch runs the conversation itself: it keeps the thread, hands the model the tools of your apps, and asks you before anything changes. Memory, skills and the browser all work with models that can call tools.

Good to know

  • Conch supplies the tools. Tool-capable models can use memory, connected apps and files in this conversation’s work folder, with the same permission checks and Undo tracking. Commands require the OS sandbox, cannot access the network and have no unrestricted fallback. Chat-only models are labelled in the picker. If a model turns out not to take tools, Conch asks it again without them and says so.
  • Spend is tracked. Conch records each turn, and you can set a budget. See Offline and at a limit.