Skip to main content

Overview

Propel is an independent AI code review layer that reviews every pull request with multi-model analysis and agentic validation. Unlike single-model reviewers or linters that operate in isolation, Propel orchestrates multiple frontier models, validates findings through an agent layer, and applies your company-specific rules and knowledge base to produce broader coverage with less noise. Propel acts as a true second opinion, separate from code generation tools.

How it works

1

Connect your repositories

Install the GitHub or GitLab app and select your repositories.Propel builds structured context across your codebase and ingests your custom rules, markdown knowledge base, and team preferences at the company, repository, and developer level.
2

Great review out of the box

Propel automatically analyzes every new pull request using multiple agents to identify and validate findings and suggested instructions for you or your agents.
3

Even better over time

Through your reactions and replies, Propel better understands the comments you want to see and do not want to see.

Explore docs

Features

Explore core Propel review capabilities and developer actions.

Configurations

Tune review behavior, rules, ignore patterns, and ticket routing.

API documentation

Integrate Propel with endpoint guides, request schemas, and examples.

What's new

See recent releases and product improvements in the changelog.