How commitplain works

A platform that converts GitHub commit activity into executive reports ready to review and send. This guide explains how the system interacts with your data, what access it requests, how generation works, and what controls exist at each step of the process.

GitHub access is handled via standard OAuth. The access token is obtained and used exclusively on the server — it is never transmitted to the browser or stored on the client.

• Only the repositories you explicitly link from the dashboard are read
• The scope requested is 'repo' for private repositories
• You can revoke access at any time from your GitHub account
• The system cannot perform any action on your repository — read-only commit access

commitplain reads your repository to generate the report — the Indie plan reads only commit metadata, the Agency plan also reads actual code diffs. In no case do we store your source code in our database: diffs are processed in memory during generation and discarded when it finishes.

1. 01

   You select period and filters

   You define the date range and the commit types to include. The system reads those commits directly from GitHub.

2. 02

   The draft is generated

   Commits are processed on the server. The draft is generated in the language and tone you have configured.

3. 03

   The draft is yours alone

   No one else sees the draft. The client is not notified. The report waits in your dashboard until you act.

4. 04

   You approve and send

   Only after your explicit approval does the system send the email to the recipient you have configured.

No report is sent to the client without a direct action from you. Automation exists — but you configure and control it. Without active automation, every send requires your manual approval.

commitplain enforces usage limits at the plan level to ensure service quality for all users. These limits are checked server-side before generation begins.

Each project can have a text description — up to 50 KB — that provides context to the AI before it generates the report. This field accepts plain text or Markdown and can be pasted directly or imported from a .txt or .md file. The context is never sent to the client. It is used only during the generation step on the server to improve the relevance and accuracy of the executive summary. Typical content: the project's business objective, the client's technical level, the domain vocabulary, and any constraints the AI should be aware of.

commitplain supports automated report generation and sending. Automation is disabled by default — you must enable it explicitly per project. When active, reports are generated on the configured schedule (daily, weekly, or monthly) and sent directly to the configured recipients without requiring manual approval. Enabling automation removes the human review gate for that project. You can disable automation at any time from the project settings. Daily automation requires the Agency plan.

commitplain stores the following data per user account:

• Repository links (name, visibility, GitHub ID)
• Generated report drafts and their approval status
• Project context text (if added)
• Report settings per project (language, tone, recipient)
• Send history (date, recipient, report content)

All user data is deleted within 30 days of account deletion. Data is stored in EU infrastructure (eu-west-1). See the Privacy Policy for full details on data handling, subprocessors, and retention schedules.