Disbug
Disbug is a browser extension and MCP-native tool that captures complete bug evidence—screenshots, console errors, network logs, DOM context, and user events—with one click and delivers it directly to AI coding agents like Claude Code, Gemini, and Cursor. The tool eliminates manual context copying by letting agents pull sessions and fetch specific data fields on demand through a CLI or MCP server, accelerating the path from bug capture to merged fix. It is free for solo developers, with paid plans for teams requiring advanced collaboration and higher usage limits.
- Developer Tools
- Artificial Intelligence
- Software Only
- AI Agents
Funding
Founders
Founder details are not available yet.
Product
Problem
AI coding agents often struggle to fix bugs because they lack the full context needed to understand what went wrong in the browser. Developers waste time manually copying screenshots, console errors, and network logs into chat prompts, which slows down the feedback loop and reduces the effectiveness of AI-assisted development.
Solution
Disbug provides a one-click browser extension that captures a complete evidence package—including screenshots, DOM selectors, console errors, failed network requests, user-event trails, browser metadata, and local storage—and delivers it directly to AI coding agents. The tool integrates with Claude Code, Gemini, Cursor, and any MCP-compatible agent through a CLI or MCP server, allowing the agent to pull the latest session, search past pins, and fetch only the fields it needs. This creates a seamless workflow from bug capture to AI-generated fix, eliminating manual context transfer and accelerating the debugging cycle. The workflow moves from browser bug to merged commit in minutes rather than hours, as agents can read the evidence, write the fix, run it, and open a pull request.
Target Audience
Primary users are solo developers and small engineering teams building with AI coding agents like Claude Code, Gemini, and Cursor who need a faster, more reliable way to feed browser bugs and debugging context into their AI workflows.
Features
- One-click Chrome extension capture that grabs screenshots, DOM selectors, console errors, failed network requests, user-event trails, browser metadata, and local storage
- MCP server integration built on the open Model Context Protocol, letting AI agents pull sessions, search past pins, and fetch specific data fields on demand
- CLI tool for command-line workflows, compatible with Claude Desktop, Cursor, Codex, OpenClaw, and Hermes
- Session management that saves captured bugs to projects for organized tracking and retrieval
- Cross-tool compatibility with all major MCP-compatible coding agents, with no lock-in to a single AI vendor
- Read-only, scoped agent access with team and project authorization controls, revocable at any time