Integrations
Use TOKNADO with the AI tools you already run
TOKNADO gives you one OpenAI-compatible endpoint backed by a managed model pool. Any tool that lets you set a custom base URL and API key can use it. Pick your tool below for copy-paste setup, a verification test, and the errors people actually hit.
How every integration works
- Sign up, then copy your API key and endpoint from the dashboard.
- In your tool, choose its OpenAI-compatible / custom provider option.
- Paste the endpoint as the base URL, paste the key, and enter a model ID.
- Run the one-line curl check from the tool's page, then start working.
Setup guides
- OpenClawAdd TOKNADO as a custom provider in the OpenClaw wizard or config.
- Hermes AgentPoint Hermes Agent at a custom OpenAI-compatible endpoint.
- Codex CLIDefine a custom model provider in ~/.codex/config.toml.
- Claude CodeRun Claude Code through an Anthropic-format layer in front of TOKNADO.
- OpenCodeAdd TOKNADO as a custom provider in opencode.json.
- Open WebUIAdd TOKNADO as an OpenAI connection in Admin Settings.
- DeepSeekUse DeepSeek-class models in OpenAI-compatible coding tools.
- ClineUse Cline's OpenAI Compatible provider with TOKNADO.
- AiderUse OPENAI_API_BASE and the openai/ model prefix.
- Roo CodeUse Roo Code's OpenAI Compatible provider, per mode if you like.
Don't see your tool? If it accepts a custom OpenAI-compatible base URL and API key, the same three values apply. Read the general setup guide.
Point your tools at TOKNADO
One OpenAI-compatible endpoint, a managed model pool behind it, and usage you can see in the Storm Console. 3-day trial, up to 250 requests per rolling hour.
- Starter $7.99/mo 2,500 requests per rolling hour
- Builder $12.99/mo 5,000 requests per rolling hour
- Studio $19.99/mo 7,500 requests per rolling hour
- Team $29.99/mo 10,000 requests per rolling hour
Start free trial See pricing Setup guide
Hourly limits apply over the previous 60 minutes. Service is best effort; upstream availability can reduce throughput.