API & MCP · early access

    API & MCP: coming soon.

    We're opening our synthetic data pipeline to code and AI agents: request labelled, QA-checked images of the scenes your detector struggles with, without waiting on a project. It isn't open yet. Join the early-access list and we'll invite you to the beta.
    Two synthetic parking-lot CCTV frames with boxes for person, weapon and car
    person · weapon · car

    01The idea

    The same pipeline as our datasets, on demand.

    See how the platform works. Until the API opens, we run generation for you on request.

    From code

    Ask for labelled CCTV-style images of a scene and scenario, with YOLO or COCO boxes, from a script or training pipeline.

    From an AI agent

    An MCP server lets assistants such as Claude or Cursor search the library and order the data a detector is missing.

    Annotate your own frames

    Send images you already have and get boxes and attributes back, checked by the same QA as our datasets.

    Need data now? Buy a library dataset, request custom images or video on your own camera views, or run the False-Positive Stress Test on the model you have.

    Questions

    Early-access questions

    Can I use the API today?

    Not yet. The API and MCP server are being prepared for a private beta. Join the early-access list and we'll invite you when it opens.

    What can I buy today?

    Library datasets with a research or commercial licence, custom image and video datasets built on your own camera views, and the False-Positive Stress Test for an existing model.

    What will it cost?

    We'll publish pricing when the beta opens. Early-access members hear first.

    Early access

    Get early access.

    Tell us what you'd build. We'll email you when the beta opens.