Build on Virgo

Virgo is an enterprise reasoning graph for engineering teams and AI agents. It captures work across your stack, the code, tickets, conversations, documents, and meetings, and connects it into one graph that every engineer and agent can query.

These docs are for developers extending that graph: connecting new systems, shaping how their data is ingested, writing retrievers, and building agents that read from it.

Start building

How Virgo fits together

Connectors read your source systems and hand Virgo a stream of records. Ingestion turns that stream into the graph: it pulls the useful entities out of the data, collapses duplicates into one canonical node, and links entities by the relationships that connect work across domains.

Retrievers read the resulting graph on behalf of a query. Virgo routes each query to the retriever built for it, and enforces your access control during retrieval, so people and their agents only ever see data they already have access to.

Each layer is extensible on its own. Add a data source, teach Virgo a new relationship, or write a retriever, and use it right away.

New to Virgo?

Start with the model the rest of these docs build on: entities, chunks, retrievers, and agents.

The Python SDK is in private preview. Email sales@virgo.fyi for early access while these docs fill in. MCP and REST API references are coming soon.