Text models
Connect a DeepSeek harness to an API endpoint
Verify that the client, model ID, and tool format work together.
Check the harness contract
Identify the API protocol the runner expects, its base URL format, and how it discovers models. Set an available ModelPort text model ID through configuration.
Validate in three passes
Start with a short response, then verify streaming, then run a tool call and return its result. Record request IDs and usage for each pass.
Example request
curl "https://api.apimodel.store/v1/models" \ -H "Authorization: Bearer $MODELPORT_API_KEY"