Potpie launched in early 2025 with a graph-first approach to codebase understanding. Before any agent ran, it indexed every file, class, function, and their relationships into a Neo4j knowledge graph using tree-sitter parsing. When an agent answered a question or generated code, it navigated this graph rather than relying on keyword search or embedding similarity — giving it structural knowledge of how the codebase fitted together: which functions called which others, which classes implemented which interfaces, and how changes in one area rippled through the system.
The platform provided purpose-built agents for specific tasks rather than a single general-purpose assistant: a Debugging Agent for root cause analysis, a Code Generation Agent for feature implementation, a Test Generation Agent for coverage improvements, and a Code Review Agent for quality analysis. Each was tuned for its specific task domain, with the graph providing the contextual navigation that made specialised agents effective on large codebases.
Potpie was notable for its Apache 2.0 licence, Docker-based self-hosting option, and VS Code extension alongside the cloud service at app.potpie.ai. The custom agent builder allowed teams to define their own agents with specific instructions and tool access, making the platform extensible for organisation-specific workflows. Potpie reached 5,500 GitHub stars with v1.1.1 released May 2026, indicating an active development community.