DOCUMENTATION

Make your first request

Use https://api.profundoai.com/v1 with your API key and any of the available models: glm-5.2, minimax-m3, or nemotron-ultra.

Send your key as Authorization: Bearer $PROFUNDO_API_KEY. The Chat Completions and Responses API endpoints accept the model IDs listed on GET /v1/models.

What's supported

Profundo supports both Chat Completions and the Responses API. Accepted Chat Completions request fields:

• model
• messages
• stream
• temperature
• max_tokens
• top_p
• frequency_penalty
• presence_penalty
• n
• stop
• seed
• user
• reasoning_effort
• response_format
• tools
• tool_choice
• parallel_tool_calls
• top_logprobs
• stream_options
• metadata
• service_tier

The API supports text input and output. Some models (including MiniMax M3) accept image input; most are text-only. No model supports audio, embeddings, transcription, file storage, assistants, batch, or fine-tuning endpoints. Tool calling, parallel tool calls, structured output (response_format), streaming, and configurable reasoning effort are fully supported where the model allows.

How pricing works

Membership is request-first with no token meter. Plus includes 250 requests per day. Priority includes 30,000 requests per month — we may raise this cap as demand grows. Fair use and service protection controls apply. There is no public SLA.

Start here