Welcome to Nectr
Nectr is an open-source AI agent that automatically reviews every pull request in plain English. Connect a repository, and when a PR is opened, Nectr posts a structured review as a GitHub comment — covering bugs, security issues, performance concerns, and style improvements — with a clear verdict and inline suggestions.Quickstart
Get up and running with Nectr in under 5 minutes
How It Works
Understand the architecture and PR review flow
API Reference
Explore the complete REST API documentation
Self-Hosting Guide
Deploy Nectr on your own infrastructure
Key Features
AI-Powered Reviews
Claude Sonnet 4.6 analyzes every PR for bugs, security vulnerabilities, performance issues, and code style with intelligent, context-aware feedback.
Knowledge Graph
Neo4j tracks file ownership, developer expertise, and PR relationships to provide contextual insights and identify subject matter experts.
Semantic Memory
Mem0 remembers project patterns, coding standards, and individual developer habits to deliver increasingly personalized reviews over time.
Live Context Integration
MCP bidirectional integration pulls live data from Linear, Sentry, and Slack to ground every review in real production context.
What Makes Nectr Different
Contextual Intelligence
Contextual Intelligence
Unlike traditional linters, Nectr understands your project’s history. It knows who owns which files, what patterns you’ve established, and what decisions you’ve made — delivering reviews that get smarter with every PR.
Production-Aware Analysis
Production-Aware Analysis
Nectr doesn’t just read code — it pulls production errors from Sentry, linked issues from Linear, and team discussions from Slack to provide reviews grounded in real-world impact.
Self-Hostable & Open Source
Self-Hostable & Open Source
Own your data and your API keys. Nectr is designed to be forked and self-hosted with full control over your infrastructure, making it ideal for teams with strict data sovereignty requirements.
Parallel Specialized Agents
Parallel Specialized Agents
Enable parallel mode to run three specialized AI agents concurrently — one for security, one for performance, and one for style — with a synthesis agent combining their insights into a unified review.
How It Works
When you open or update a pull request:GitHub Webhook Triggers
GitHub sends a webhook event to Nectr’s FastAPI backend, which validates the signature and queues the review in the background.
Context Gathering
Nectr pulls the PR diff, fetches related issues from Linear, production errors from Sentry, and team discussions from Slack to build comprehensive context.
Knowledge Graph Query
Neo4j identifies file experts (developers who frequently touch the changed files) and related past PRs to inform the review.
AI Analysis
Claude Sonnet 4.6 analyzes the code with full context, checking for security vulnerabilities, performance issues, bugs, and style violations.
Review Posted
Nectr posts a structured review comment on GitHub with inline suggestions, a summary, and a clear verdict (APPROVE, REQUEST_CHANGES, or COMMENT).
Tech Stack
FastAPI
High-performance Python backend with async support
Next.js 15
React 19 frontend with server components
Neo4j
Knowledge graph for file ownership and PR relationships
Mem0
Semantic memory for project patterns and developer habits
Claude Sonnet 4.6
State-of-the-art AI model for code analysis
PostgreSQL
Relational database for users, repos, and reviews
Getting Started
Quickstart Guide
Connect your first repository and get your first AI review in minutes
Dashboard Overview
Explore team analytics, review history, and contributor metrics