Verify with a minimal request whether OmniDino is connected correctly.
Verify the endpoint URL, API key, model name, and protocol endpoint first, then troubleshoot the client itself, such as Codex, Claude Code, Gemini CLI, Cherry Studio, or Dify.
Select an endpoint and terminal to generate a test command automatically.
This page does not transmit your API key. Commands are generated only in your current browser, so do not paste a real key on a public device.
Select an endpoint and click “Generate Command.”Validate each endpoint step by step.
Start with the Models endpoint, then test the actual model endpoint. This quickly reveals whether the issue is with the gateway, key, model, or protocol layer.
Validate the Gateway and API Key
This is the most basic check. It confirms that the domain is reachable, the request headers are formatted correctly, and the API key passes basic authentication.
curl https://api.omnidino.com/v1/models \
-H "Authorization: Bearer YOUR_API_KEY"Chat Completions
Commonly used by standard OpenAI-compatible clients, Cherry Studio, Open WebUI, Dify, and most chat models.
curl https://api.omnidino.com/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_MODEL",
"messages": [
{
"role": "user",
"content": "Reply with: OmniDino connected"
}
]
}'Responses
Commonly used by Codex, OpenAI Responses-compatible clients, agents, and selected reasoning models.
curl https://api.omnidino.com/v1/responses \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_MODEL",
"input": "Reply with: OmniDino connected"
}'Claude Messages
Used by Claude Code, Claude Desktop, and Anthropic-compatible tools. Clients usually use the root URL and append the specific request path themselves.
curl https://api.omnidino.com/v1/messages \
-H "x-api-key: YOUR_API_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_CLAUDE_MODEL",
"max_tokens": 64,
"messages": [
{
"role": "user",
"content": "Reply with: OmniDino connected"
}
]
}'Gemini Routing
Gemini models in OmniDino may be exposed through an OpenAI-compatible protocol or through the native Gemini protocol.
curl https://api.omnidino.com/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_GEMINI_MODEL",
"messages": [
{
"role": "user",
"content": "Reply with: OmniDino connected"
}
]
}'Image Generation
Used for image models that support the OpenAI Images-compatible endpoint, including image-generation models configured in the OmniDino model catalog.
curl https://api.omnidino.com/v1/images/generations \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_IMAGE_MODEL",
"prompt": "A small coding dinosaur in a clean futuristic office",
"size": "1024x1024",
"quality": "high"
}'Locate issues quickly by status code.
The same error can have several causes. Check the high-probability items listed here first, then review the OmniDino dashboard and client logs.
Unauthorized
The request did not pass API key authentication.
- Copy the complete API key again and check for leading or trailing spaces.
- Confirm that the key has not been disabled, deleted, or expired.
- Do not add the Bearer prefix twice in a field that only asks for the key.
Insufficient Balance or Quota
The account balance, token quota, or group limits are insufficient to complete the request.
- Check the current account balance.
- Check the quota limit assigned to this API key.
- Review the model price, group multiplier, and actual request cost.
Not Found
The request path, HTTP method, or protocol is usually mismatched.
- Check whether GET is being used on an endpoint that only supports POST.
- Check whether the Claude root URL and OpenAI /v1 URL have been mixed up.
- Confirm that the current channel supports the Responses, Messages, or Images path.
Too Many Requests
The request rate, concurrency, or upstream rate limit has been reached.
- Reduce concurrency and retry frequency.
- Add exponential backoff to automatic retries.
- Check whether multiple clients are sharing the same key.
Service or Upstream Error
The issue may originate from the model upstream, channel configuration, protocol conversion, or reverse proxy.
- Test another model to determine whether the issue affects only one model.
- Check the OmniDino status page and NewAPI logs.
- Confirm that the channel type matches the upstream protocol.
Model Not Found or Access Denied
The model name, channel binding, token group, or endpoint type may be incorrect.
- Copy the exact model name from the model catalog.
- Confirm that the model has been added to the relevant channel.
- Confirm that the current token group can access the model.
Endpoint Not Implemented
This commonly occurs when a Responses request is routed through an incompatible channel or converter.
- Check whether Wire API is set to Responses.
- Confirm that the channel type is not incorrectly set to NewAPI or another conversion mode.
- Confirm that the upstream itself supports the target endpoint.
Connection Successful
The gateway accepted the request and returned a valid result.
- Continue checking whether the client reads the response correctly.
- Review the usage record and balance change in the dashboard.
- Set a sensible quota and a dedicated API key for production use.
Do not reinstall the client first.
Work through the following layers in order to distinguish network, key, balance, model, and protocol issues more quickly.
When Is the Connection Considered Successful?
When all conditions below are met, the OmniDino connection and target model call are working correctly.
Provide complete context when contacting support.
Copy the template below, complete it, and send it to [email protected]. Hide your complete API key.
Account Email: Operating System: Client Name: Client Version: Base URL： Endpoint Type: Model Name: HTTP Status Code: Full Error Message: Time of Issue: Can /v1/models be accessed successfully? Is sufficient balance available? Troubleshooting steps already tried: Note: Do not include passwords, complete API keys, payment card information, or other sensitive data.