Provider overview

How Cerevisor connects to models, provider types, the library, resolution order, and per-agent overrides.

The seven provider types

Type What it is Authentication Cost model
Anthropic Direct API to Claude (Opus, Sonnet, Haiku). Cerevisor runs the agents itself. API key. Per token.
Gemini Direct API to Google's Gemini models (2.5 / 3.x families). Cerevisor runs the agents itself. API key from Google AI Studio. Per token, with a >200K input-tier bump on Pro models.
Claude Agent SDK Anthropic's official agent SDK (Claude Code's engine) running in-process. Claude Code runs the work; Cerevisor supervises. Anthropic API key (separate keychain slot from the Anthropic provider) or a Claude Pro/Max subscription. Per token via your Anthropic account, or your Claude plan's quota (shown as $0).
OpenAI-compatible Any endpoint speaking the OpenAI Chat Completions API. API key (or none, for Ollama). Per token, or free (local).
OpenAI Codex CLI Sidecar to OpenAI's codex command-line tool. Codex runs the work; Cerevisor supervises. Browser sign-in with your ChatGPT account (device code as fallback). Subscription quota (shown as $0).
Cursor Agent In-process Cursor SDK runtime. Cursor runs the work; Cerevisor supervises. Cursor API key. Per token, billed to your Cursor account.
Antigravity Sidecar to Google's Antigravity CLI (agy). Antigravity runs the work; Cerevisor supervises. Multi-model: Gemini, Claude, and an open-weight model on one Google plan. Browser sign-in with your Google account (agy login). Your Google plan; Antigravity reports no usage counts, so Analytics shows no figure.

Anthropic, Gemini, Claude Agent SDK, Codex CLI, Cursor Agent, and Antigravity are singletons: one entry per type. OpenAI-compatible allows N entries, so you can have OpenAI + Ollama + OpenRouter side by side as separate library entries.

Who runs the agent. With Anthropic, Gemini, and OpenAI-compatible providers, Cerevisor runs the agents itself — per-tool permission prompts, mid-run steering, skill injection, and contract retries all work. With Claude Agent SDK, Codex CLI, Cursor Agent, and Antigravity, the external tool runs the work in its own engine while Cerevisor supervises: you set the rules up front, and you trade those step-by-step Cerevisor features for that tool's own behavior, tool catalog, and (usually) subscription billing. Cerevisor shows you exactly this trade-off on the provider's capability card when you add it.

The Provider Library

All your credentials live in the Provider Library, accessed via Settings → Providers.

Each entry has:

  • A name (auto-generated from type + base URL, editable).
  • The provider type.
  • Connection details (base URL, model defaults).
  • A "default" badge: exactly one entry is the default at any time.

API keys are stored in the OS keychain. They're never written to a .cerevisor file and never logged.

Provider resolution

When an agent runs, Cerevisor resolves which provider to use via this chain:

  1. Agent-level override. If agent.providerId is set, use that.
  2. Workflow-level override. Else, if workflow.settings.providerId is set, use that.
  3. Library default. Else, use the library's default-marked entry.
  4. Fallback. Else, the first enabled credential.

This means a workflow can run mixed providers (different agents use different providers) without you having to configure them per workflow.

The default

One library entry is always marked default. New workflows use it. Agents in those workflows inherit it unless you override.

Change the default in Settings → Providers → Set as default on the entry you want.

What background features use the default

Cerevisor has a few features that use the default provider quietly in the background:

  • Chat builder: uses the default unless overridden in Settings → Chat Builder.
  • Run retrospectives: at end-of-run, Cerevisor summarizes the run for memory storage. Uses the default.
  • Meta-cognition enrichment: periodic summary of what Cerevisor has learned about you. Uses the default.
  • Skill summarization: the LLM-rich descriptions of skills (the upgrade pass). Uses the default.
  • Memory crystallization: periodic distillation of your freeform memory entries.

These features run on your default even when the default is Codex, Cursor, Claude Code, or Antigravity: each call runs as one small agent task on that subscription, so a subscription-only setup keeps all background features working. If the default fails at runtime (not signed in, quota exhausted), Cerevisor automatically retries the call on another chat-capable credential in your library — no toggle to flip. Any spend from such a fallback is recorded and shows up in Analytics → Models, so you always know what a fallback cost.

Workflow execution vs. background features

The provider system is split into two paths:

Path Routing
Workflow execution Per-agent resolution. Mixed providers OK. All seven types OK.
Background features The default first — including subscription providers — with automatic fallback to another chat-capable credential if the default fails.

This is important because it lets you build a workflow where, say, every agent runs on Codex or Claude Agent SDK (billed to a subscription you already pay for) while background features quietly keep working too.

Where to go next

Back to docs