Other Providers¶

Beyond the big three, VidaiMock ships drop-in support for the other surfaces real applications hit. All are bundled and require zero configuration.

Azure OpenAI¶

POST /openai/deployments/{deployment}/chat/completions
POST /openai/deployments/{deployment}/embeddings

Azure uses the OpenAI request/response shape under a deployment-scoped path. The bundled Azure provider matches the deployment path pattern and serves the OpenAI templates, so any Azure OpenAI SDK works unchanged — just point its endpoint at VidaiMock.

AWS Bedrock¶

POST /model/{model_id}/invoke
POST /model/{model_id}/converse
POST /model/{model_id}/invoke-with-response-stream
POST /model/{model_id}/converse-stream

Bedrock's binary event-stream framing is emulated for the streaming variants.

Vertex AI¶

POST /v1/projects/{project}/locations/{location}/publishers/google/models/{model}:generateContent

The Gemini response shape under Vertex's project/location-scoped path.

OpenAI-compatible providers¶

Cohere, Mistral, and Groq all expose OpenAI-compatible chat endpoints. Point any OpenAI SDK at VidaiMock with the relevant model name — the OpenAI-compatible catch-all handles /v1/chat/completions for these providers, including streaming.

Anything else¶

Every endpoint above is just a bundled provider YAML + Tera template. To add a provider VidaiMock does not ship, drop a YAML into your --config-dir with a matching regex and a response template. See Provider config reference and Overriding bundled defaults.