The WildLens AI Manifesto
A digital field guide for the real world.
Most modern mobile apps are engineered to maximize your screen time through continuous notifications, social feeds, and gamified streaks. WildLens AI is built on the opposite philosophy:
“Identify it. Learn something. Look up. Go explore.”
Open-Weight AI: imageomics/bioclip-2
Unlike generic cloud vision APIs that treat living organisms as commercial catalog objects, WildLens AI utilizes BioCLIP-2 (Vision Transformer ViT-L/14), an open-weight foundation model developed by Imageomics and Ohio State University, trained on the massive TreeOfLife-200M dataset comprising over 200 million biological specimens across global taxa.
ArchitectureOpenCLIP ViT-L/14768-dim Visual-Taxonomic Embeddings
Taxonomic GroundingTreeOfLife-10M10,000,000 biological plates
LicenseMIT License100% Permissive Open Source
100% Local Processing & Privacy Architecture
Every photo you take and every note you record stays strictly on your device. Zero external cloud API calls, zero telemetry tracking, and zero advertising SDKs.
photo_camera1. Device CameraLocal photo sensor
memory2. Local BioCLIPFastAPI / CPU inference
database3. IndexedDBEncrypted browser store
cloud_off4. Zero Cloud UplinkWorks offline on trail
shieldModel Limitations & Safety Guidelines
- Not for Foraging or Edibility: Never consume wild plants or mushrooms based on AI predictions alone. Chemical toxicity and macroscopic spores require specialized laboratory mycological analysis.
- Anatomical Visual Correspondence: AI classification reflects visual morphological similarity in photographs, not genetic sequencing or guaranteed biological certainty.
- Respect Wildlife: Keep a respectful distance from all wild animals and nesting birds. Do not handle or disturb specimens for the sake of a photograph.