Rafflesia
Deterministic biological primitives for life-science research agents.
Models on Rafflesia
Services
Homology search
activeDeterministic ranked protein-sequence candidate search against an immutable database release. It uses the same scientific contract as POST /v1/homology while adding a durable retained run, provenance, billing, cancellation, and lifecycle events.
Standalone MSA
previewBuild a deterministic, content-addressed, unpaired protein MSA with the current or an explicitly pinned generic configuration.
AlphaFold 2 MSA
previewBuild the full AlphaFold 2 UniRef90, BFD plus UniRef30, and MGnify MSA through immutable Tomato releases.
AlphaFold-Multimer MSA
previewBuild one AlphaFold-Multimer component MSA with full AF2 unpaired lanes and an immutable UniProt pairing-source lane.
AlphaFold 3 MSA
previewBuild AlphaFold 3 UniRef90, Small BFD, MGnify, and UniProt MSA inputs through immutable Tomato releases.
Boltz-2 MSA
previewBuild complete per-chain Boltz key-sequence CSV inputs and a ready-to-use ZIP bundle from immutable Tomato releases.
Chai-1 MSA
previewBuild per-chain Chai aligned-Parquet inputs and a ready-to-use ZIP bundle from immutable Tomato releases.
ColabFold MSA
previewBuild ColabFold-compatible UniRef30 and environmental query-anchored alignments through immutable Tomato releases.
Generic MSA pairing
previewBuild one generic MSA per distinct protein component and assemble deterministic cross-component row assignments without a predictor-specific compatibility claim.
About
Rafflesia builds deterministic, provenance-carrying biological primitives that research agents can call directly: homology search, multiple sequence alignment construction, structure prediction, and embedding services, each returning typed evidence with stable schemas and exit codes.
Its homology stack is Tomato, an embedding-era search engine that combines ESM2 residue embeddings, object-native approximate nearest neighbor retrieval, exact MaxSim late interaction, and embedding-scored Smith-Waterman alignment against immutable database releases.
The first-party endpoints below are Rafflesia services. The hosted models elsewhere in this catalog are open-weight models published by their respective labs and served on Rafflesia infrastructure via the runs API.
Use these models
Every endpoint speaks the same run lifecycle: POST the endpoint id and a typed input, get a run back with outputs, provenance, and measured billing.
# one run lifecycle, every endpoint
curl https://api.rafflesia.ai/v1/runs \
-H "Authorization: Bearer $RAFFLESIA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"endpoint": "rafflesia/homology",
"input": { "sequence": "MKTAYIAK..." }
}'