Skip to main content

The Memory Graph: Exploring What Revo Knows

The Graph tab in Memory shows the people, companies, projects and commitments Revo has learned about, and how they connect. Filter it, search it, and click any fact back to the meeting or document it came from.

Overview

Revo builds up knowledge about your business from every meeting, document and issue in your workspace. The Memory Graph is the picture of that knowledge — the people, companies, projects, products and commitments Revo has learned about, and the relationships between them.

It's the answer to two questions people ask constantly:

  • "What does Revo actually know about us?"

  • "Where did it learn that?"

Every fact in the graph traces back to a source you can open.


Where to find it

Click Memory in the left sidebar. You'll see two tabs:

  • Graph — the visual map of what Revo knows. This is where Memory opens by default.

  • Modules — the Intelligence Modules view, covered in Intelligence Modules.

The Graph tab appears once your workspace has the brain enabled. If you only see the Modules page, the graph isn't switched on for your workspace yet — get in touch and we'll take a look.


Reading the graph

Each node is something Revo knows about — a person, a company, a project, a product, a commitment. Each line between two nodes is a relationship Revo has established, like "works at", "owns", or "committed to".

  • Brighter nodes are better established. The more Revo has seen something confirmed across different sources, the more it stands out.

  • Lines carry their own confidence. A relationship Revo has seen corroborated in several places is drawn more strongly than one it has seen once.

  • Hover any node for a quick summary without leaving the canvas.

  • Pan and zoom to move around. Use the fit and spacing controls to reset the view or spread out a dense cluster.

The graph follows your app theme, so it re-inks itself in dark mode.


Filtering what you see

Open the filters panel to narrow the graph down. Changes apply as you make them — there's no Apply button to hunt for.

  • Entity types — show or hide people, companies, projects and so on. Each type shows how many of it there are, and you can partially select within a type.

  • Isolated entities — things Revo knows about that aren't connected to anything else yet. Hide them to clean up the view, or show them to spot knowledge that hasn't been linked up.

  • Ended relationships — connections that are no longer current, like someone who has left a company. Hidden by default; turn them on to see history.

  • Since — limit the graph to what Revo has learned within a date range.

The legend explains what each colour and shape means.


Two ways to search

The search bar sits above the canvas and does two different things depending on how you use it.

1. Type to filter. As you type, everything that doesn't match dims away, leaving the matches lit. This is a live text match — good for "where is Acme in here?"

2. Press Enter to ask. Hitting Enter runs a proper search across the graph and returns a written answer with citations. Good for questions like "what did we commit to Acme?" or "who owns the migration project?" Each citation is clickable and takes you to the source.


Entity and relationship details

Click any node or line to open its detail panel.

For an entity, you'll see:

  • Aliases — the other names this thing goes by. Useful for spotting when Revo has learned the same company under two names.

  • Unique identifiers — email addresses, domains and the like, which are how Revo tells two similarly-named things apart.

  • Relationships — everything it's connected to. Hovering one highlights it out on the canvas so you can see where it sits.

  • Sources — where Revo learned about it.

For a relationship, you'll see:

  • Provenance — what Revo saw that made it draw this connection.

  • Supersession chain — what this fact replaced. When something changes (a project moves to a new owner, a date shifts), Revo keeps the old version and links it, so you can see how the picture evolved rather than just the current state.


Clicking back to the source

This is the part worth knowing about. Sources in the detail panels are links. Click one and you land on the meeting, document or issue the fact came from.

So if Revo believes a deadline is the 15th, you can go and read the sentence in the transcript where someone said so. Nothing in the graph is a black box.

Occasionally a source can't be resolved to a link — usually because it came from something outside your workspace. In that case you'll see the reference without a clickable link.


Coming back to the graph

The graph is cached for your session. Leave Memory, do something else, come back, and it renders instantly in the same layout you left it in — no reloading and no waiting for it to settle again.

Reload the page when you want to pull in the latest.


What you can see

The graph respects the same permissions as the rest of Revo. You see knowledge drawn from sources you already have access to — a private meeting you weren't part of doesn't contribute anything you can see, and nothing in the graph exposes content you couldn't open directly.


Frequently Asked Questions

Something in the graph is wrong. Can I fix it?

Start by clicking through to the source — often the graph is faithfully reflecting something that was said inaccurately in a meeting, or a document that's now out of date. Correcting or archiving the source is the durable fix, since Revo rebuilds from your workspace content. If a fact looks wrong and the source doesn't explain it, let us know.

The same company appears twice under different names

Open each one and compare the aliases and unique identifiers in the detail panel. Revo separates entities when it doesn't have enough evidence that they're the same thing — usually more mentions across your sources will merge them over time.

Why is my graph mostly empty?

The graph is built from your workspace content, so a new workspace starts sparse. Recording meetings and connecting your document sources fills it in quickly. Also check your filters — hiding isolated entities on a young workspace can hide most of what's there.

What's the difference between the Graph and Intelligence Modules?

They're two views of the same underlying knowledge. Modules organise what Revo knows by topic — Company DNA, Product Features, and so on — and are what feed your email drafts. The Graph shows the raw structure: the individual things Revo knows about and how they connect. Use Modules to check coverage, and the Graph to trace a specific fact.

Does the graph use credits?

Browsing, filtering and text-matching in the graph are free. Pressing Enter to ask a question runs an AI search and uses credits like any other Revo query. See How Credits Work.

Did this answer your question?