Modal
ActiveAn AI agent for Modal. It works in a scene with a live connection to Modal. It asks before it writes, and every run is traced.
Run Python on demand with arbitrary GPUs (H100, A100, L4), deploy as web endpoints, schedule cron, mount volumes — all via decorator-based Python. Daslab's only 'manufactory' provider: agent-authored code becomes a durable, callable service.
What you can connect
Add these to your scene and AI gets access.
A Modal serverless GPU/compute workspace.
Connect Modal over MCP
Daslab is itself an MCP server.
Connect Modal in a scene, then pin that scene's URL in Claude Code, Cursor, or any MCP client. The agent there gets your Modal tools, and your approval gate and trace travel with them.
https://daslab.run/mcp/<your-world>/<your-scene>
One scene can hold Modal and everything else you connect, so a single endpoint carries them all. Set up a client.
Use cases
- Deploy custom GPU inference endpoints from agent-authored Python
- Run scheduled batch jobs (cron) against your data with autoscaling
- Spin up sandboxes for ad-hoc compute; persistent volumes for state
From Modal
Similar integrations
Others in Infrastructure.
Ready to try Modal with Daslab?
Get started with CLI