actively fundraising · seed + series a
why this is worth building.
model choice becomes computed, cited artifacts and a checkable record; we are looking for partners who want to widen access to that workflow.
the thesis
the compute got cheap; access to it did not
- access
the gap between powerful ai models and the scientists who need them is vast. our mission is to bridge that gap — not by building more models, but by making existing breakthroughs accessible to everyone, regardless of technical background.
- give back
we're building toward becoming a public benefit corporation — committed to supporting the open-source ecosystem we depend on. we allocate revenue to maintainers, offer equity grants to contributors of critical dependencies, and fund fellowships for scientific software developers.
- enterprise
the compliance page records what is implemented, in progress, and planned. current enterprise work includes auditability, electronic signatures, and deployment controls.
see the compliance status - attribution
we don't pretend to invent the models we run. esmfold, diffdock, admet-ai, aizynthfinder — that is other people's science, and every result we return names its source. what we built is the layer around them: the runtime that plans and executes a run, the tamper-evident ledger that records it, and the passports that turn raw model output into a pharma-ready deliverable. we acknowledge our debt to rdkit, mol*, biopython, alphafold, and hundreds of maintainers.
platform depth
a run hands back something a team can use.
- compute
GPU + CPU serverless functions spanning the public workflow domains — zero infrastructure management for end users.
- reports
computational results are aggregated into structured reports — compound, target, safety, literature, and therapeutic-area deliverables.
outputs range from CSV and molecular files to publication-ready PDFs, including journal-ready source files for direct submission.
- visualization
molecular structures, docking poses, genomic circular plots, dose-response curves and survival analyses — rendered server-side at publication resolution, embedded in reports or exported as standalone figures.
- pharmacology
numerical workflows for dose-response modeling, principal component analysis, spectral analysis, differential equation solving, and signal processing — through natural language, without specialized software licenses.
business model & licensing
free to start, plans or prepaid credit packs after, FSL-licensed throughout.
our platform uses the FSL-1.1-Apache-2.0 (functional source license), allowing free use for non-competing purposes and converting to Apache 2.0 after two years. this builds community trust while preventing direct competitors from repackaging our work. enterprise features are separately licensed.
- free tier
500 credits a month — full access to the platform's scientific computing workflows.
- paid tiers
plans that scale from individual researchers to teams, plus prepaid credit packs.
- enterprise
custom licensing for compliance features (21 cfr part 11, sso, white-label).
- vs incumbents
accessible pricing against legacy enterprise vendors ($50k+/year for incumbents).
interested in investing?
let's talk.
this page contains forward-looking statements about our vision and market opportunity. contact our team for detailed financial information, the pitch deck, and due diligence materials.