OntoCast as a service — document-to-knowledge-graph pipelines without the infrastructure.
The managed version of our engine stack: documents in, a governed knowledge graph out — downloadable as RDF, queryable with SPARQL and scriptable through an API. Delivery into your graph store and agent access over MCP are planned.
Early access: our team approves new accounts before their first extraction.
What you get
Managed extraction, end to end
Managed extraction pipeline
Send documents from the console or the API; get ontology-grounded facts back as RDF. Queueing, retries and scaling are handled for you.
Ontology management
Bring your own ontology and SHACL shapes, or let OntoCast draft an ontology from your documents. Runs can extend it, and every extraction stays consistent with it.
Export, query, and (planned) delivery
Download each run or a whole project as Turtle, N-Triples, N-Quads, TriG or JSON-LD, or query it with SPARQL. Delivery straight into Neo4j, ArangoDB or Postgres is planned.
API-first
Everything scriptable with an API key: estimate a run, submit it, follow its progress, pull the results. Built for pipelines, not dashboards-only workflows.
Costs and access under control
Every run shows an estimate first and stops at a cost limit you can set. Queries are read-only, with caps on run time, result size and rate. Serving the graph to agents over MCP is planned.
Same engine, hosted
Built on code you can inspect
The platform runs OntoCast — the same code that's on GitHub and PyPI. Evaluate the engine yourself, then let us run it at scale.
Getting started
A 30-day Implementation Sprint
One domain, one set of sources, success criteria agreed in advance — and a working, governed graph at the end. Fixed fee, fixed scope.
Week 1
Ingestion & profiling
Connect your sources, profile the data, draft the model.
Week 2
Schema refinement
Review the contract together — identities, edges, evolution plan.
Week 3
Engine deployment
Materialize into the graph store you already operate.
Week 4
Evaluation
Verify against your success criteria; hand over the contract.
FAQ
Questions we get
What does it cost?+
Usage-based: a run pays a fee per thousand characters of text, a small fee per PDF page or image it converts (plain text, Markdown, HTML and Office files pay none), and the model tokens it uses, at the prices shown in the console. Before each run the console shows an estimate and the most it may cost, and no run costs more than its limit. A run that ends in an error is not charged; a run you cancel, or that reaches its limit, pays only the tokens it used. New accounts get a starter credit when approved; during early access, more credit is added by request. Accounts start on the starter plan; a pro plan for larger projects is available on request. Early-access partners get direct input into the roadmap.
How is my data handled?+
Your documents are processed for extraction only and are not used to train models. Data-processing agreements are available for early-access partners.
Can I self-host instead?+
Yes — the platform runs the same engines you can deploy yourself today: OntoCast and GraFlo are both open source (Apache-2.0). The platform adds hosting, scaling and management.
Can I use it now?+
Yes, in early access: sign in to the console and our team approves new accounts before their first extraction. General availability comes later; join the list below for updates.
Early access
Get platform updates
Updates on the platform as it grows. Unsubscribe anytime.
Request a demo
Tell us about your use case
A short call, a walkthrough on your kind of data, and an honest answer on fit.
Or email us directly: team@growgraph.dev

