TrailNote: an offline photo guesser for birds and plants

Description and use

TrailNote is an offline tool that uses CLIP to guess birds and plants from photos, providing top 3 predictions with confidence scores and optional logging.

Problem

Users want to identify birds or plants during walks without relying on online services or API keys.

Outcome

TrailNote delivers quick, offline bird and plant identification with confidence scores, enabling field users to get instant guesses without internet access.

Workflow

• Take a photo of a bird or plant
• Run the TrailNote CLI command with the image path
• The tool loads the CLIP model from local cache
• It scores the image against ~35 hardcoded labels
• Outputs top 3 guesses with confidence percentages
• Optionally logs the guess with date to a JSON file
• Optionally upload image via Gradio page

Evidence

TrailNote is a small command-line tool for walks. Take a photo of a bird or plant, run one command, and it gives you its top 3 guesses with confidence scores.

Source

The model is OpenAI's CLIP (clip-vit-base-patch32), an open-weight model, run through the Hugging Face transformers zero-shot image classification pipeline.

Source

On my CPU a run took about 7 seconds. PyTorch alone is around 750MB.

Source

It runs on your own machine with no API key and no account.

Source