Biological foundation models for target identification and biomarker discovery, run and fine-tuned from Claude, Codex or any MCP client.
Generative pretrained transformer trained on 33 million human cells for single-cell annotation, batch correction, and perturbation prediction.
Single-cell foundation model pretrained on about 30 million human transcriptomes, using rank-value encoding for context-aware gene network inference.
Genomic foundation model built on the Hyena operator, processing DNA at single-nucleotide resolution with context windows up to 1 million tokens.
Genomic foundation model trained on 9.3 trillion DNA base pairs across all domains of life, with 40B parameters and a 1-million-token context.
Single-cell foundation model producing species-agnostic cell embeddings by representing genes through frozen ESM-2 protein language model embeddings.
Biological foundation models are used and post-trained on customer data through the Helical platform, driven directly from Codex, Claude or any other MCP client. A request described in the client uploads data to secured per-project storage, launches the computation job and returns the results, with each job held for explicit approval. The Helical console handles credits, files and the account. The open-source Python package needs no account and runs on a local GPU, under per-model licences.
Billing is consumption based, in credits at one credit to the US dollar. Each job's credit quote is shown before it runs, and execution is approved from the MCP client. A run costs a fixed amount per job plus its cell count times a per-model, per-cell rate; the rates are listed in the console behind a sign-in. Failed or cancelled runs are refunded automatically, and larger credit purchases and the Academic Partner Program add free credits. The open-source package is free.
Last verified