Prelint

Product Introduction

1. Definition: Prelint is an AI-powered product review and specification enforcement platform for software development teams. It operates as a specialized static analysis tool that integrates directly into the Git workflow (GitHub/GitLab) to analyze pull requests against a company's documented product intent, business rules, and architectural decisions.
2. Core Value Proposition: Prelint exists to prevent product drift in AI-generated and human-written code by automatically enforcing product specifications. Its primary function is to catch misalignment between code changes and documented business logic, compliance requirements, and strategic decisions before they are merged and shipped, thereby acting as a guardrail for AI coding agents and developers.

Main Features

1. Automated Product Intent Review: Prelint automatically reviews every pull request by parsing and understanding product specification documents (Markdown, YAML) stored in the repository. It uses natural language processing (NLP) and semantic analysis to compare code diffs against constraints defined in Architecture Decision Records (ADRs), product specs, and compliance docs. It flags contradictions in real-time within the PR interface.
2. Machine-Readable Constraint Detection: The platform identifies and surfaces conflicts between specifications, such as a new feature violating a documented pricing rule or introducing a non-standardized tool. It provides inline comments with specific references to conflicting documents (e.g., specs/pricing/settlements.md), enabling immediate, context-aware corrections.
3. Specification-Based Q&A for Stakeholders: Prelint includes a feature that allows non-engineers (e.g., product managers, sales) to ask questions about product functionality and receive answers grounded directly in the version-controlled specifications. This reduces interruptions for engineers and eliminates "Slack archaeology" by providing a single source of truth.

Problems Solved

1. Pain Point: "Product Drift" or "Specification Drift," where AI coding assistants or developers inadvertently write code that deviates from established product requirements, business logic, or compliance rules because they lack full context. This leads to costly bugs, security gaps, revenue leakage, and wasted engineering effort.
2. Target Audience: Engineering teams using AI coding tools (e.g., GitHub Copilot, Cursor, v0), Engineering Managers, Product Managers, and Compliance Officers in tech companies. It is particularly critical for fast-moving startups and scale-ups leveraging AI to accelerate development.
3. Use Cases: Catching foreign exchange (FX) risk in multi-currency billing code; preventing GDPR violations from new data storage; enforcing tooling standards to avoid vendor sprawl; ensuring new features align with the product roadmap; answering stakeholder questions about product decisions without blocking engineers.

Unique Advantages

1. Differentiation: Unlike traditional code review tools (e.g., SonarQube, Greptile) that focus on code quality, security vulnerabilities, and syntax, Prelint operates at the product and business logic layer. It complements code review by answering "Should this code exist?" rather than just "Does this code work?".
2. Key Innovation: Prelint's approach of treating product specifications as machine-readable, version-controlled sources of truth and integrating their enforcement directly into the CI/CD pipeline. This creates a closed feedback loop where AI agents can self-correct based on spec violations before human review, significantly bounding error accumulation.

Frequently Asked Questions (FAQ)

1. How does Prelint differ from GitHub's built-in code review or Copilot? GitHub's native tools and Copilot assist with code generation and peer review of implementation quality. Prelint is a separate, specialized layer that reviews the intent of the code against your business rules and product specifications. It ensures the generated code aligns with what the product should do, not just how it does it technically.
2. Is my source code and proprietary business data safe with Prelint? Yes, Prelint employs a tenant-isolated infrastructure model. Each customer's data runs in isolated containers and storage. The company has a strict zero-retention policy for raw source code and does not use customer data to train its AI models. All data is encrypted in transit (TLS 1.3) and at rest (AES-256).
3. What formats does Prelint support for product specifications? Prelint primarily supports human-readable, text-based formats like Markdown and YAML. These documents are stored directly in your Git repository alongside your code, making them version-controlled, diff-able, and accessible to both humans and the Prelint analysis engine.
4. Can Prelint work without AI-generated code? Absolutely. While it is exceptionally effective at catching drift from AI coding assistants, Prelint also reviews pull requests from human developers. It serves as an automated checkpoint to ensure all code changes, regardless of source, adhere to the documented product specs and business constraints.
5. How is pricing structured for Prelint? Prelint uses a consumption-based pricing model of $1 per completed review, with no per-seat subscription fees. New users receive $10 in free credits to start. Open-source projects can use Prelint for free, and a startup program is available.