Skip to main content
Perceptron Mk1.5 understands images, video, and audio, grounds answers in space and time, and calls functions that your application provides. Use the same chat completions API for questions, transcription, structured extraction, object tracking, and tool-assisted workflows.

Create an API key

Get your key from the Perceptron platform

Join Discord

Get help and see what others are building

Make your first request

Install the Perceptron Python SDK and set your API key:
Use the SDK’s image(), video(), and audio() helpers for local paths, bytes, URLs, or uploaded file references. Local paths and bytes are encoded inline; use Files to upload once and reuse a file ID. See Python setup for a complete integration.

Explore the capabilities

Image understanding

Ask questions, read text, and locate objects.

Video understanding

Describe actions and locate events in time.

Video tracking

Follow objects with timestamped geometry inside track tags.

Tool calling

Connect the model to functions executed by your application.

Audio understanding

Summarize recordings, ask questions, transcribe speech, and locate events in time.

Reasoning

Choose reasoning effort and budget for the final answer.

Multiple assets

Compare media and map annotations with asset_idx.

Structured outputs

Constrain final answers with JSON Schema or regex.
See the model card for limits and pricing. For production integrations, read Tokenization, Error messages, and Scaling.