video
Find the Moment
Drop a long recording, ask a question: a vision agent searches it coarse-to-fine, cuts the clip, subtitles it in three languages, optional Omni remix.
Demos
Apps built on GMI Cloud
Each demo runs on the same API and compute you build with. Open one, then read the docs.
Community demos
Built by the GMI Cloud community and GMI DevRel. 49 demos, each one a small app that calls GMI inference and returns a result.
video
Physics Twin
Footage in, physics out: Gemini 4 Argon tracks a moving object across 48 frames, writes and runs the fit, and puts its g next to 9.81 with an error bar.
video
Lookout
A security camera wall: six-second clips go to Nemotron 3.5 Super VL, which returns events, positions and rule-breaking alerts.
video
Narrator
Audio description for blind viewers: MiniMax M3.1-Flash reads the frames, writes one line per silent gap, the browser speaks it in the gap. WebVTT out.
video
Hermes × GMI MCP × Wan 3.0
Hermes Agent on the GMI MCP server: one prompt, six tool calls, one Wan 3.0 clip. The three livestream clips, the prompt, and the preflight runbook.