Roboflow
ActiveAn AI agent for Roboflow. It works in a scene with a live connection to Roboflow. It asks before it writes, and every run is traced.
Roboflow gives agents a vision modality: hosted inference for object detection, classification, and segmentation; chained Workflows that combine detection, OCR, and classification; and Universe — 100k+ public models you can call without training. Pair it with iOS camera/photos for capture and the staged-effects layer for human-approved actions on visual triggers.
What you can connect
Add these to your scene and AI gets access.
Roboflow workspace credentials for computer vision
A Roboflow project — dataset + trained model versions
A pinned public model from Roboflow Universe
Connect Roboflow over MCP
Daslab is itself an MCP server.
Connect Roboflow in a scene, then pin that scene's URL in Claude Code, Cursor, or any MCP client. The agent there gets your Roboflow tools, and your approval gate and trace travel with them.
https://daslab.run/mcp/<your-world>/<your-scene>One scene can hold Roboflow and everything else you connect, so a single endpoint carries them all. Set up a client.
Use cases
- Detect manufacturing defects in line photos and stage a Slack alert + Linear ticket
- Audit retail shelf compliance from store photos and email the brand manager
- Triage drone footage by class (vehicles, people, infrastructure) into a morning briefing
- Run an active-learning loop — agent surfaces low-confidence predictions for human labeling
From Roboflow
Similar integrations
Others in AI & ML.
Ready to try Roboflow with Daslab?
Get started with CLI