System Diagnostics & Local Storage

Settings & AI Telemetry

Transparent hardware metrics and local storage controls for WildLens AI.

memoryOpen-Weight Foundation Model

Checking...

Vision Model (BioCLIP-2)imageomics/bioclip-2 (ViT-L/14)License: MIT · TreeOfLife-200M

Curated Knowledge Basetaxonomy_kb.py (Deterministic)Zero LLM · Mandatory Mushroom Guard

Vision Execution Devicecpu (Multi-threaded)

Taxa Grounding & Offline36 Reference Taxa · 100% Edge Verified

verifiedBoth Models 100% Open-Weight & MIT / Apache-2.0 Licensed

dnsFastAPI Backend Server & Cloud Host

Disconnected

Architecture Notice: WildLens AI is deployed as a static web app on GitHub Pages. Because GitHub Pages does not run Python or PyTorch, live BioCLIP-2 vision AI identification runs in an external FastAPI server.

You can host the FastAPI server on Render, Railway, Hugging Face Spaces, or run it on your local network/machine.

FastAPI Service Base URLDefault Fallback

Deployment Quick Guide:

• Render / Railway: Deploy backend/Dockerfile or uvicorn app.main:app.

• CORS Setup: Set CORS_ORIGINS=https://suguda-thakur-marndi.github.io in backend environment variables.

• GitHub Actions: In repo Settings > Secrets and variables > Actions > Variables, set NEXT_PUBLIC_API_URL so every build automatically points to your production backend.

folder_sharedLocal Journal Data & Retention

IndexedDB Record Count0 observation dossiers saved in browser storage sandbox

Clear all observations from local storage

videocamCamera Permission Guidance

When in the field, mobile browsers may request camera access to capture specimens. Your video stream is evaluated in temporary memory frame-by-frame and is never recorded or streamed to any remote server. If camera access is denied, you can always upload or select photos from your device gallery.

WildLens AI v1.4.0 · Hacktoberfest 2026 EditionMIT License · Pure Naturalist Instrument