Rastrum
Identify any living thing — anywhere
Rastrum exists to make every living thing identifiable by anyone, anywhere — even without cell signal, even without formal training, even in the languages that field guides forgot.
How It Works
Observe
Capture photos, audio, or video of any species in the field — online or offline.
Identify
Our AI pipeline analyzes your evidence using PlantNet, BirdNET, and Claude Vision with regional accuracy.
Contribute
Validated observations join a growing open dataset exported to GBIF and CONABIO for real conservation.
Why Rastrum
Offline-First
Works without cell signal. Regional ML models cached on-device via ONNX + WebAssembly.
Multi-Modal ID
Photo, audio, video, and ecological evidence — tracks, scat, burrows — all in one platform.
Indigenous Languages
Zapoteco, Mixteco, Náhuatl, Maya, Tsotsil/Tseltal. Built with CARE principles and community consent.
LATAM-Focused
Trained on megadiverse Neotropical data, not just North American and European species.
Open Source
AGPL-3.0 server, Apache 2.0 SDKs. Every observation strengthens the commons.
Conservation Pipeline
Direct exports to GBIF, CONABIO, CONANP, and regional herbaria across Latin America.