evroc Think Models (Shared) API

OpenAI-compatible API for shared models.

Version: 1.0

Contact information:
API Support
support@evroc.dev

/v1/audio/transcriptions

POST

Summary:

Create audio transcription

Description:

Transcribes audio into text using the OpenAI audio transcription API. This endpoint accepts multipart/form-data and proxies requests to the model backend. Required: file (audio file), model (string). Handler-processed optional fields: language (string), response_format (string), prompt (string), temperature (float).

Parameters
NameLocated inDescriptionRequiredSchema
AuthorizationheaderBearer token for API key authenticationYesstring
Responses
CodeDescription
200Transcription response in OpenAI format with text and usage statistics
400Bad request - Invalid form data, missing file, or missing model field
404Unknown model
500Internal server error or model configuration issue
502Failed to connect to model backend

/v1/chat/completions

POST

Summary:

Create chat completion

Description:

Creates a chat completion response for the given conversation. This endpoint is OpenAI-compatible and proxies requests to the model backend. Required: model (string). Handler-processed optional fields: stream (bool), stream_options.include_usage (bool). All other OpenAI chat completion parameters (messages, temperature, max_tokens, top_p, etc.) are passed through to the backend.

Parameters
NameLocated inDescriptionRequiredSchema
AuthorizationheaderBearer token for API key authenticationYesstring
Responses
CodeDescription
200Chat completion response in OpenAI format with id, object, created, model, choices array, and usage statistics
400Bad request - Invalid JSON or missing model field
404Unknown model
500Internal server error or model configuration issue
502Failed to connect to model backend

/v1/completions

POST

Summary:

Create text completion

Description:

Creates a completion for the provided prompt using the legacy OpenAI completions API. This endpoint proxies requests to the model backend. Required: model (string), prompt (string/array). Handler-processed optional fields: stream (bool), stream_options.include_usage (bool). Additional parameters are defined by openai.CompletionNewParams and passed to the backend.

Parameters
NameLocated inDescriptionRequiredSchema
AuthorizationheaderBearer token for API key authenticationYesstring
Responses
CodeDescription
200Completion response in OpenAI format
400Bad request - Invalid JSON or missing required fields
404Unknown model
500Internal server error or model configuration issue
502Failed to connect to model backend

/v1/models

GET

Summary:

List available models

Description:

Fetches and aggregates models from all configured backend endpoints. Returns models in OpenAI format with model ID replaced by alias and owned_by set to 'evroc think platform'. Includes max_model_len when available from backend.

Parameters
NameLocated inDescriptionRequiredSchema
AuthorizationheaderBearer token for API key authenticationYesstring
Responses
CodeDescription
200List of available models in OpenAI pagination format with object='list' and data array containing model objects (id, object, created, owned_by, max_model_len)
500Internal server error - No models configured or no models available