Sirca
The context layer for AI.
Connect your company's sources and Sirca builds a living knowledge graph of everything your company knows: every person, project, deal, and document, and how they all connect. Explore it. Ask it questions in plain language. Serve it to every AI application through one API.
Connect
Google Drive, Slack, Salesforce, Postgres, email, and the systems your company already runs on.
Resolve
People, companies, projects, products, deals, and documents become one coherent graph.
Serve
Your team explores it directly, and every AI application queries one structured, sourced API.
Ingestion
Connect a source. Watch it build.
Connect Google Drive, or Slack, Salesforce, Postgres, email, and Sirca starts reading. Within minutes, entities appear: the people, companies, projects, and products inside your data, with the relationships between them mapped automatically.
You're watching the structure of your company assemble itself from your own systems. Connect a second source and the graph gets sharper: the same people and projects resolve across systems into one coherent picture.
Exploration
Explore everything your company knows
Entity Explorer
Browse the graph two ways. A filterable, sortable list: by type, source, confidence, date. Or a visual network view: click a node, see its connections, expand outward to trace how anything relates to anything else.
Entity detail
Click any entity and see every property it has, the exact source each one came from, down to the document and page, its confidence score, and when it was observed. See every relationship, grouped and dated. See the full history of how the entity changed over time. Merge, split, or correct anything by hand.
Nothing is a black box. Every fact traces to its source. Always.
Ask
Ask in plain language
Type a question the way you'd ask a colleague and get a structured answer assembled across every connected system, with citations back to the source of each fact. No query language. No guessing which document to open.
Example question
"Who owns Project Phoenix, what's its status, and what's blocking it?"
Answer shape
Owner, current status, blockers, next actions, and cited source records.
Each fact points back to the document, message, record, or page it came from.
Trust
Built to be trusted
Source-traceable
Every answer carries its provenance.
Contradiction-aware
When two sources disagree, Sirca surfaces both and lets you resolve it, rather than silently picking one.
Permission-aware
Access is enforced at the graph level. Sirca inherits the permissions of your source systems, so a user only ever sees facts they are already allowed to see.
Always current
When a source changes, the graph updates. The picture is never stale.
A salary pulled from an HR doc stays invisible to anyone without HR access.
Use
Two ways to use it
The platform
Your team logs in to explore the graph, ask questions, and understand the company in a way no search tool allows. Built like infrastructure: fast, dense, quiet. A tool you keep open all day.
The API
Every AI application you build, from chatbots and agents to analytics and automation, queries one endpoint for structured, sourced, permissioned company knowledge, instead of each team rebuilding its own brittle retrieval. One context layer underneath everything.
Connectors
Connectors
More connectors are added continuously.
Coverage
One platform, any company
Your data defines the domain, not us. A mining company connects geological and financial data; a hospital connects workforce and capacity data; a software team connects its tools.
Sirca does the same thing every time: ingest, resolve, relate, serve. Already running in production at scale, with the same engine generalising to any company that has data and needs its AI to actually understand it.