Getting Started
Grove builds a continuously updated, conflict-resolved understanding of your company — delivered as files any agent can read. This guide covers the core concepts and gets you set up with your first data source.
What is Grove?
Grove is the context layer that turns your company's scattered data into a brain every agent already knows how to read. Instead of searching at query time, Grove continuously synthesizes signal from your existing tools — Slack, email, CRM, docs, meetings — and surfaces the result as structured files on disk.
Any agent that reads files can consume Grove's output without custom integration. No connectors, no per-agent setup. Just a filesystem that already knows your company, updated daily.
How Grove is different from RAG
Retrieval-augmented generation (RAG) searches for relevant content when a question is asked. Grove maintains a continuously updated representation of your company before anyone asks.
| Capability | RAG | Grove |
|---|---|---|
| When context is built | At query time | Continuously |
| Conflict resolution | None | Source-ranked |
| Identity unification | No | Yes |
| Staleness awareness | No | Yes |
| Agent compatibility | Custom per agent | Any file reader |
| Compounding value | Resets every query | Grows daily |
Quickstart: connect your first data source
The fastest way to get started with Grove is the company interview. It takes about 15 minutes and produces a first-pass context graph from a structured conversation about how your company actually works.
- 01
Start the interview
Navigate to /interview. Grove's AI will ask you 6–8 focused questions about your team, tools, processes, and informal knowledge.
- 02
Review your context graph
After the interview, Grove produces a structured snapshot of what it learned — company name, key processes, decision patterns, tools, and open questions.
- 03
Connect live data sources
Once you have a working baseline, connect Slack, Google Workspace, email, or CRM to keep the graph continuously updated.
Understanding the context graph
The context graph is Grove's core output: a synthesized, source-tracked representation of your company. It resolves conflicts between sources, unifies identity across fragmented mentions, tracks freshness, and surfaces cross-source inferences that no single document contains.
The graph is surfaced as structured files on disk. Every agent that reads files — Claude Code, Cursor, OpenClaw, or any custom agent — can consume it without custom integration. The files are updated daily as new signals come in from your connected sources.
Next
Connecting Sources — Coming soon