Setting Up ProdE

Setup itself takes just a few minutes. Indexing runs in the background and takes longer (see below for what to expect).

1. Create your account: sign up at app.prode.ai/auth using Google sign-in or with your work email and a password. If you register with email, complete the verification email sent to your inbox.
2. Connect your Git provider: link GitHub, Bitbucket, GitLab, or Azure DevOps.
3. Select the repositories you want indexed: after your provider is connected, ProdE shows the repositories available in your workspaces or organizations. Pick the ones you want indexed.
4. Start asking ProdE questions: once indexing finishes, open the chat in ProdE. See ProdE Chat for the kinds of questions you can ask.

After these four steps, ProdE is ready to use. The next two are optional but highly recommended; they extend ProdE into your editor and Slack.

5. Connect your coding agent (optional but highly recommended): bring ProdE's knowledge into Cursor, VS Code, Windsurf, or Claude Code via MCP.
6. Connect Slack (optional but highly recommended): create a Slack application and register it with ProdE. You can then ask ProdE questions directly from Slack.

What to expect during indexing

ProdE indexes your repositories in two passes:

1. Call graph pass. ProdE builds the structural map of your code: files, symbols, calls, and cross-repo relationships. This pass is fast. Even a 200,000-line codebase typically finishes in a matter of minutes.
2. LLM pass. ProdE enriches each component with semantic understanding (what each piece does, why it exists, how it fits into the product). This pass takes longer and scales with the size of the repository. As a rough guide, plan for around 20 minutes for 200,000 lines of code, and around 1 hour for 1 million lines.

You can start using ProdE for basic structural questions as soon as the call graph pass completes. The LLM pass continues in the background, and answers become richer as it progresses.