Floumy Documentation

Code and AI

Connect repositories, review engineering metrics, use AI-assisted content generation, and configure the MCP server.

Enable Code First

The Code area only appears after you enable code integration in Project Settings.

Once enabled, Floumy checks whether the project is already connected to:

  • GitHub
  • GitLab

If a connection exists, Floumy sends you straight to the matching integration page.

GitHub

GitHub setup is OAuth-based.

After connecting, you can:

  • Select the repository for the current project
  • Change the selected repository later
  • Disconnect the repository
  • Review pull requests
  • Review delivery metrics

GitHub metrics

Floumy shows chart-based metrics for:

  • Pull request cycle time
  • Average merge time
  • First review time

GitLab

GitLab setup starts with a personal access token.

The setup screen expects:

  • A token with API scope
  • Maintainer role on the GitLab side

After saving the token, you select a GitLab project and get the same core view:

  • Merge requests
  • Cycle time
  • Merge time
  • First review time

You can later change the selected GitLab project or disconnect the integration.

AI Features in Floumy

Floumy includes focused AI helpers inside the workflow rather than a separate idea-generation area.

Current AI-assisted actions include:

  • Generate key results from an objective
  • Generate initiative descriptions
  • Generate work-item descriptions
  • Generate initiatives from a request

These helpers work best when your project description and artifact titles are specific.

MCP Server

The user sidebar can expose an MCP Server page for organizations with that feature enabled.

That page gives you:

  • A personal Floumy MCP token
  • Setup instructions for supported clients
  • A token refresh action

Supported setup guides in the UI

  • Claude for Desktop
  • Cursor
  • Windsurf
  • Zed

Refreshing the token rotates your credentials. If you refresh it, update any local client configuration that still uses the old token.

Best Practices

  1. Keep one Floumy project aligned to one repository or repository group with a clear delivery boundary.
  2. Add a good project description before relying on AI-generated drafts.
  3. Refresh your MCP token if you suspect it has been exposed.
  4. Use PR and MR metrics as trend signals, not as a substitute for engineering judgment.