Introduction
CertaEdge provides a unified model API endpoint for applications and developer tools. Create an API Key, choose a model available to your account, and send requests using its supported protocol.
The three connection settings
| Setting | What to enter for CertaEdge |
|---|---|
| Base URL | Use https://certaedge.com/v1 for OpenAI-compatible APIs. |
| API Key | A model API key created in the console. |
| Model | Copy the complete model ID from the model catalog. |
Tools handle paths differently. Claude Code uses https://certaedge.com and appends /v1/messages. OpenAI-compatible SDKs usually need the address that includes /v1.
Start with your first request
- Create an account and prepare an available balance.
- Create an API Key, with quota and permissions suited to your application.
- Confirm the Base URL and model.
- Send a short test request.
- Find your client in the tool guides and follow its setup instructions.
Choose the matching protocol
CertaEdge offers several compatible APIs. The model, endpoint, and parameters must match. A tool with a custom Base URL setting does not necessarily support every model capability.
| Protocol | Common endpoint |
|---|---|
| OpenAI Chat Completions | /v1/chat/completions |
| OpenAI Responses | /v1/responses |
| Anthropic Messages | /v1/messages |
| Gemini native API | /v1beta/models/{model}:generateContent |
Refer to the current model catalog and your client's supported features.
API Keys and usage
- Create separate API Keys for different applications to track usage, restrict access, and disable them independently.
- Do not commit real keys to a code repository or include them in public screenshots.
- After a request, open Usage Logs to check the model, status, and cost.
- A CertaEdge API Key and a third-party client subscription are separate services. The client's own terms determine access to its paid features.
If a request fails, start with FAQ and Troubleshooting.
For available video, music, or image-generation services, choose an endpoint in Generation Task APIs. These usually follow a submit-then-query flow and do not use a standard chat request body.