heddle Review

heddle is an open-source AI coding agent runtime tool developed by Jay / Fienna Liang that enables developers, software engineers, and development teams to inspect live repositories, make bounded code changes, and maintain continuity across coding sessions. It functions as a local-first terminal CLI with persistent memory, sessions, approvals, heartbeat tasks, and a browser control plane. Beyond its coding agent capabilities, heddle also serves as a 'shared memory layer' for teams, integrating with existing workflows and tools such as Linear, Notion, Slack, Figma, and GitHub. This dual functionality aims to structure decisions, link them to their source, and ensure context and intent are preserved and updated across various platforms and team members. The platform is designed to be observable, with every step visible and every action traceable, emphasizing a human-in-the-loop approach for critical updates.

heddle provides a robust set of features designed to enhance the capabilities of AI coding agents and team collaboration. Its local-first architecture ensures data privacy and control, while its open-source licensing promotes transparency and community contributions. The platform's core functionalities revolve around enabling persistent, observable, and human-approved AI interactions within development workflows.

heddle is primarily targeted at individuals and teams engaged in software development and collaborative knowledge management who require a structured and traceable approach to AI agent integration. Its design addresses the need for continuity, oversight, and shared context in complex project environments.

heddle operates on a freemium business model, offering core functionality without charge. Specific details regarding paid tiers or advanced enterprise features are not publicly available. The platform is currently engaging early access partners who are past personal AI, have a real team workflow, and are willing to provide structured feedback through monthly working sessions and async issues for at least a quarter. This suggests a focus on refining the product based on user input before a comprehensive pricing structure is released.