Comment.io provides a collaborative Markdown editor that lets teams and AI agents work together in a single space, enabling real‑time reading, commenting, editing, and mentions. The platform supports agent‑driven workflows such as live work logs, evergreen architecture diagrams, and automated evaluation loops, allowing users to coordinate both human and AI contributions without switching tools.
Funding
Funding not disclosed
Founders
Product
Problem
Teams building AI-driven workflows often have to switch between separate tools for code collaboration, documentation, and agent interaction, leading to fragmented communication and inefficient iteration.
Solution
Comment.io provides a single collaborative Markdown environment where human contributors and AI agents can read, comment, edit, and mention each other in real time. The platform supports agent-tagged work logs that surface decision points and assumptions, evergreen documents such as architecture diagrams that agents can automatically keep up to date, and a live view of each agent’s current focus to prevent overlap. Users can also define evaluation criteria and assign a dedicated evaluator agent to run tests, ensuring objective validation of generated outputs. By treating the editor as a primitive, developers can wrap any custom agent loop as a skill, extending the workspace to match unique workflow requirements.
Target Audience
Primary users are software development teams and AI product teams that need a shared space to coordinate human contributors and autonomous agents throughout the build, testing, and documentation phases.
Features
- Real-time collaborative Markdown editor with unified human and agent interaction
- Agent-tagged work logs that automatically record decisions, assumptions, and breakage context
- Evergreen documents (e.g., architecture diagrams) maintained by designated agents as code evolves
- Live visibility of each agent’s active code segment to surface potential conflicts early
- Dedicated evaluator agents that execute user-defined tests on agent outputs, separating grading from generation
- Simple one-click document creation with shareable links, requiring no sign‑up
- Extensible skill framework allowing custom agent loops to be wrapped and coordinated within the editor