BioCLIP 2

Model

Open model that identifies plant, animal and fungus species in photos offline

Price: Free, open weights
Access: None, runs locally

About

BioCLIP 2 is an MIT-licensed CLIP model from the Imageomics Institute, trained on TreeOfLife-200M (214M images, 952K taxa), for zero-shot species classification across the tree of life. Run it on your own machine with pybioclip; accuracy is lower for rarely photographed taxa.

What you can do with it

Predict the species of an animal, plant or fungus in a photo on your own server
Classify organism photos against your own list of candidate species
Create image embeddings of organisms for search or a custom classifier

Get started

Install the package with pip install pybioclip
Run bioclip predict image.jpg, or use TreeOfLifeClassifier in Python

Example

from bioclip import Rank
from bioclip.predict import TreeOfLifeClassifier

classifier = TreeOfLifeClassifier()  # BioCLIP 2 by default
predictions = classifier.predict("Ursus-arctos.jpeg", Rank.SPECIES)
for prediction in predictions:
    print(prediction["species"], "-", prediction["score"])

Details

Hosting: Self-hosted, Runs locally
Available in: Worldwide
Official SDKs: Python
MCP server: None