sokrypton/ColabFold

Health v2 · maintenance only

Adoption not part of the score

2889 stars · 742 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

ColabFold is a set of tools and Google Colab notebooks that make protein structure prediction with AlphaFold2, AlphaFold3, ESMFold, and related models accessible to everyone. It provides both browser-based notebooks and a local Python package for predicting protein structures and complexes.

Use cases

• predict protein 3D structure from amino acid sequence
• model protein-protein complexes
• run alphafold2 without a local GPU
• batch predict many protein structures
• generate multiple sequence alignments with mmseqs2
• predict structures locally on my own machine

When to choose

• you need protein structure prediction without setting up AlphaFold yourself
• you want free GPU-based prediction via Google Colab
• you want a pip-installable wrapper for local AlphaFold-style predictions

When to avoid

• you need full control over AlphaFold internals or custom training
• you cannot use Google Colab and lack a capable GPU locally
• you need guaranteed privacy for proprietary sequences in the cloud notebooks

Facets

application · maturity active

machine-learning deep-learning sdk bioinformatics machine-learning artificial-intelligence python cross-platform protein-folding alphafold structure-prediction google-colab jupyter-notebooks mmseqs2 computational-biology web-server gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sokrypton/ColabFold")