LLM Registry — Client Setup Guides
The LLM Registry lets your team point any AI runtime (Codex CLI, Claude Code, Cursor, a Python SDK, plain curl) at the Cequence AI Gateway instead of calling OpenAI, Anthropic, or Google directly. The gateway takes care of authentication, provider credentials, usage attribution, and enforcement — the runtime doesn't need to know which provider is on the other side.
This section documents the runtime / provider / mode combinations that have been verified end-to-end.
Quick start
Pick your runtime from the sidebar, then choose the provider and mode inside the guide:
- curl (Generic HTTP) — OpenAI, OpenRouter, Anthropic, Gemini, Bedrock
- Codex CLI — OpenAI, OpenRouter
- Gemini CLI — Gemini (Direct and via Agent Persona)
- Cursor IDE — OpenRouter, Bedrock, Gemini (Passthrough); Anthropic, OpenAI (partial)
- Claude Code — Anthropic (coming soon)
- Python SDKs — OpenAI SDK, Anthropic SDK, Google Gen AI SDK (coming soon)
Terminology
The same three words show up on every guide. If they're new to you, read this once and the rest of the section is straightforward.
Provider. The upstream LLM service the entry points at — OpenAI, Anthropic, OpenRouter, Google Gemini, Amazon Bedrock, and so on. Set when the entry is created.
Mode. How the provider credential is handled:
- Gateway-managed key — you upload the provider's API key once when you create the entry. Your runtime doesn't see it; it only sends its Agent Access Key.
- Passthrough — the gateway holds no provider credential. Your runtime supplies the provider key at call time, alongside its Agent Access Key. Useful for personal accounts or short-lived tokens.
Agent Access Key. The gateway's own credential, generated on the LLM Registry entry's Connect flow. Every call to the gateway must carry one — it's what identifies the caller for enforcement and audit. It's generated once and revealed only at creation time, so copy it somewhere safe.
Agent Persona. A tenant-defined scope that bundles a set of allowed tools and (optionally) LLM access under one endpoint. When the entry is reached "via an Agent Persona," the same Agent Access Key covers both the persona's tools and its LLM calls.
Verified combinations
The matrix below tracks what has been tested. ✅ means end-to-end verified against the deployed gateway.
| Runtime | OpenAI | OpenRouter | Anthropic | Gemini | Bedrock |
|---|---|---|---|---|---|
| curl (Generic HTTP) | ✅ | ✅ | ✅ | ✅ | ✅ |
| Codex CLI | ✅ | ✅ | — | — | — |
| Gemini CLI | — | — | — | ✅ | — |
| Cursor IDE | ⚠️ | ✅ | ⚠️ | ✅ | ✅ |
| Claude Code | — | — | soon | — | soon |
| Python SDKs | soon | soon | soon | soon | soon |
Each ✅ covers Direct (gateway-managed or passthrough) unless the guide notes otherwise. Via Agent Persona is covered where verified — see the mode tabs on each runtime's page.
Tips
- Test the endpoint before the client. Every runtime hides errors differently. If a client reports a generic auth failure, run a plain curl against the same base URL with your Agent Access Key to confirm the endpoint and key are correct — then move to the client's config.
- Regenerate a lost key. Agent Access Key plaintext is shown only once. If you lose it, generate a new one from the LLM Registry entry's Connect flow; there's no way to recover the original.
- One key per attach point. An Agent Access Key attaches to either an LLM Registry entry or an Agent Persona (not both). Generate a separate key for each surface a runtime reaches.
Cequence AI Gateway