You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Scigantic builds a platform for working with public scientific data: a searchable catalog of thousands of archives, notebooks that mount them directly with no download step, an in-notebook AI assistant with schema-aware context, and fine-tuning pipelines for biological sequence models. scigantic.com
The packages in this org are the query layer underneath that platform, released standalone. Each one is a permissively licensed Python client for a specific public archive or file format, self-contained and usable with no dependency on the rest of Scigantic. Together they're a library of libraries: point-lookups, search, and streaming access over public scientific data, without mirroring a database yourself or downloading more than you need. Using one on its own is a preview of the access layer that Scigantic's contextualized archives and fine-tuning infrastructure are built on top of.
SureChEMBL patent chemistry: 31M compounds and 45M patents by id, name, SMILES, InChIKey or structure search, the patents for any ChEMBL or PubChem compound, full patent text, EBI's bulk parquet read in place with DuckDB, and joins into scigantic-chembl and scigantic-bindingdb
Search all 14,000+ Brain Image Library datasets (6 PB of BICCN brain microscopy) and read light-sheet, fMOST and STPT volumes over HTTP: one slice, one thumbnail, or an OME-Zarr level, nothing downloaded
A PubChemPy-inspired PUG REST client with throttle-aware retry, caching, BioAssay/gene/protein/Tox21 data, live cross-references into ChEMBL, BindingDB and the PDB, and a local parquet mirror for b…
SureChEMBL patent chemistry from Python: compound lookup, structure search, patents by compound, Solr full-text search, full patent documents, and EBI's bulk parquet read in place with DuckDB. No mirror, no download, no key.
A PubChemPy-inspired PUG REST client with throttle-aware retry, caching, BioAssay/gene/protein/Tox21 data, live cross-references into ChEMBL, BindingDB and the PDB, and a local parquet mirror for bulk CID lookups.
MCP server for the Scigantic catalog of public scientific data archives. Search thousands of datasets and get LLM-ready schema cards without downloading anything.