Potpie is an open-source platform that creates AI agents specialized in analyzing and interacting with codebases. It targets software developers and teams looking to automate code understanding, debugging, testing, and feature development tasks. Potpie builds a knowledge graph of code to enable agents to understand complex relationships, offering pre-built agents and custom agent creation capabilities.

How It Works

Potpie constructs a knowledge graph of the codebase, capturing relationships between code components. This graph serves as the foundation for specialized AI agents that can query and reason about the code. Agents leverage a set of tools for interacting with the knowledge graph (e.g., vector similarity searches, retrieving code snippets by node ID or name) and performing actions like change detection or retrieving file structures. This approach allows for deep code understanding beyond simple text analysis.

Quick Start & Requirements

• Installation: Requires Docker, Git, and Python 3.10.x.
• Setup: Clone the repository, initialize submodules (git submodule update --init), navigate to potpie-ui, update .env with database credentials and LLM provider keys, build the UI (pnpm build), and start services (./start.sh or ./start.ps1).
• Resources: Requires PostgreSQL, Neo4j, and Redis. LLM inference and chat models need to be configured.
• Documentation: App, Documentation, API Reference.

Highlighted Details

• Offers pre-built agents for debugging, Q&A, code change analysis, integration testing, unit testing, low-level design, and code generation.
• Provides VSCode and Slack integrations for seamless workflow access.
• Supports custom agent creation via prompts and tool integration.
• Enables API access for CI/CD and automated workflows.

Maintenance & Community

• Active development with contributions welcomed.
• Links to documentation and API access are provided.

Licensing & Compatibility

• Licensed under Apache 2.0.
• Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

The setup process involves multiple services (Docker, PostgreSQL, Neo4j, Redis) and requires specific LLM configurations, which may present a moderate barrier to entry. The README indicates custom agent creation is an "Upgrade" feature, suggesting some advanced functionalities might be behind a paid tier.