Can I use an OpenAI-compatible endpoint that is not OpenAI?

Yes. Add a provider with your own API base — any service that speaks the OpenAI chat completions format works, including Azure-style gateways and self-hosted servers.

Where is the API key stored?

In FastClaw's database, as provider configuration. From the CLI you can point it at an environment variable with --api-key-env instead of pasting the key.

Integration · Models

Use OpenAI GPT models with FastClaw

Run FastClaw agents on OpenAI models — set a deployment-wide default, override per agent, and expose your agents through an OpenAI-compatible API.

Get started free See pricing

Built-in preset

The openai preset fills in https://api.openai.com/v1 and reads OPENAI_API_KEY — paste a key and pick a model.

Per-agent models

Keep a cheap default for most agents and give one agent a stronger model. Agent-scope settings shadow the system ones by name.

OpenAI-compatible in and out

Agents are served at /v1/chat/completions, so any OpenAI SDK can talk to a FastClaw agent — tools, memory and sandbox included.

Resilient turns

Timeouts and transient errors are retried with backoff, and replies cut off at the output limit are retried in smaller steps.

What you can do

Make GPT models the default for every agent on your deployment, or only for the agents that need them.
Switch an agent's model at any time from the console or with /model in fastclaw chat.
Call your agents from apps that already use the OpenAI SDK — point the base URL at FastClaw and keep your code. See the runtime API.

Set it up

Open Console → Models, click Add provider, choose OpenAI and paste your API key.
Pick a default model. To give a single agent a different model, open the agent and go to its Models tab.
Or do it in one command from the CLI:

export OPENAI_API_KEY=sk-...
fastclaw agents init alpha \
  --provider openai \
  --model openai/gpt-4o-mini \
  --api-key-env OPENAI_API_KEY

Notes

OpenAI is one of several model providers. Many teams mix it with Anthropic or route through OpenRouter for broader model choice. Building a product on top? Read FastClaw for SaaS builders.

Frequently asked questions

More Integrations

All Integrations

Anthropic

Run FastClaw agents on Claude through Anthropic's native Messages API — with prompt caching, extended thinking and image input preserved.

DeepSeek

Run FastClaw agents on DeepSeek — a built-in preset, thinking-mode support and per-agent model overrides for low-cost, capable agents.

Ollama

Point FastClaw at Ollama to run AI agents on local, open-weight models — no API key, no data leaving your machine.

OpenRouter

Connect FastClaw to OpenRouter and give every agent access to models from many vendors with a single API key.

Try it on your own agents

Start free in FastClaw Cloud, or self-host the single binary in a minute.

Get started free See pricing