Skip to main content
This walks through the shortest path to a real result: give Erdo some data, then ask an agent to do something with it.

1. Create an account

Sign up at erdo.ai. You’ll land in your organization’s workspace — your team shares it, with permissions controlling who can see and do what.

2. Add some data

Open Data and choose how to bring data in:
  • Upload a dataset — drop in a CSV or Excel file to get going immediately.
  • Connect a source — under Connectors, link a database, Google Workspace, Slack, or another tool. Sign in and authorize; the connection is private to you and encrypted.
Connecting Google Calendar here is also what enables an agent to book meetings on a voice call later.

3. Brief an agent

Start a thread and ask for what you want in plain language. Be specific about the outcome — the agent figures out the steps.
“From my signups dataset, show new signups by plan for last month and flag anything that looks off.”
The agent connects to your data, does the analysis, and shows the result — tables, charts, and its reasoning — right in the thread.

4. Turn it into something durable

From the same thread, keep going:
  • Build a page — “Turn this into a dashboard I can share with the team.” The agent produces a live page wired to your data.
  • Automate it — have the agent re-run on a schedule so the report refreshes on its own. See Activity.

5. Stay in control

As agents work, they’ll propose things — knowledge to remember, or actions that change data or reach outside Erdo. You’ll see these in Review or as an approval prompt. Nothing consequential happens without your sign-off. See Review and approvals.

Use Erdo outside the app

You don’t have to work in the web app, and you don’t have to write code:
  • MCP server — connect Claude, Cursor, or any MCP client to your data
  • CLI — drive Erdo from your terminal or an AI coding session
Building software on Erdo? The REST API and the TypeScript / Python SDKs are in the Developers tab.