PrelintDocs

Product review for
every pull request

Code review checks whether your code works. Prelint checks whether your code matches what the team agreed to build. It reads your product specs, architecture decisions, and context docs, then reviews every PR for drift, contradictions, and business logic errors before they ship.

The gap in your review process

Every engineering team has some version of this workflow: product decisions get documented somewhere (Notion, ADRs, markdown in the repo), code gets written, and a pull request gets reviewed. The review catches bugs, security issues, and style problems. What it doesn't catch: whether the code actually matches what was decided.

That gap gets worse with AI coding agents. They write correct, passing, well-structured code that violates your product specs. They don't know your pricing model, your compliance constraints, or the architectural decisions your team made six months ago.

Prelint closes that gap. It sits alongside your code review tool and checks the product layer: does this PR match the specs? Does it contradict a documented decision? Does the business logic actually do what the requirements say?

Two minutes to your first review

1

Install the GitHub App

Add Prelint to your GitHub organization and select which repositories to review. No CI changes, no YAML, no build step.

2

Add context to your repo

Drop product specs, decision logs, or architecture docs into your repository. Markdown files are indexed automatically.

3

Open a pull request

Prelint reviews every PR within seconds. Findings appear as inline comments pinned to the exact lines that triggered them.

What gets reviewed

Prelint focuses exclusively on the product layer. It doesn't duplicate your linter, your type checker, or your code review tool.

Prelint reviewsOther tools handle
Spec violations and product driftCode style and formatting
Architecture decision complianceBug detection and security scanning
Business logic correctnessType safety and test coverage
Scope drift and unplanned changesPerformance and refactoring suggestions

Guides

Choose the one that matches where you are:

Your specs become the review criteria

Prelint reads markdown files in your repository and uses them as review context. Architecture decision records, product requirements, API contracts, compliance rules. Anything you document becomes something Prelint enforces.

Context is assembled within an 80K character budget, prioritized from most specific to most general: project level specs first, then repository level specs, then general documentation. If your repo has a clone on EFS, the review agent can also search and read your full codebase to verify claims before flagging them.

The more you document, the sharper the reviews get. Teams that maintain specs and decision logs see 2.75x more accurate reviews than those running without context.

Your code stays yours

Repository clones are encrypted at rest, stored on isolated per-tenant volumes, and mounted only during reviews. Each review runs in a dedicated container that is destroyed on completion. LLM calls use zero retention APIs. No other tenant can access your code, and Prelint never uses it for training.

See the security page for full details on data handling, infrastructure isolation, and compliance.