Skip to main content
This guide gets you from installation to a result you can evaluate: a comparison of three competitors, with sources and open questions. No email sender or CRM setup is needed for this example.

Before you start

Have an OpenHive account, an internet connection, and the websites of three companies you want to compare. Agent work uses your available model access and credits; check the plan badge if a run cannot start.
1

Install and sign in

Download OpenHive from open-hive.com and follow the installation instructions for your platform.Open the app and follow the sign-in screen. If you use email-code sign-in, enter the code from your mailbox. Use a password instead is available for an account with a password.
2

Let setup finish

The first launch prepares the local runtime and downloads dependencies. Keep the app open and connected while setup runs. Use Show details on the startup screen to see what it is doing.Setup may also run after an update. If it reports an error, follow startup troubleshooting.
3

Describe a small task

Choose New Chat in the sidebar. Paste the prompt below, replace the three bracketed websites, and submit it.
OpenHive home screen showing the task composer above featured prompts.

Use the home-screen composer to state the task and the output you want.

Follow the work

OpenHive routes the request to a queen and opens a colony. Read the progress in chat. If it needs access to your browser, follow Browser automation; if it asks a question, answer in the same colony.

Check your result

A successful first run gives you three comparable entries, source links, and clearly marked unknowns. Open a few sources to check that the comparison supports its conclusions. A stopped spinner alone does not establish that the task is complete. If the output needs work, continue in the same colony:
Results may appear in chat, files, or the Data tab, depending on what the agent produces. Ask explicitly for a saved table when you want tabular output you can export.

Next

Find and qualify prospects, learn the colony workspace, or improve your next request with Write a useful task brief.