Krometrail
Debugging support for C/C++ applications via GDB or lldb-dap using the Debug Adapter Protocol.
Generates Cypress test scaffolds from recorded browser sessions for reproduction and testing.
Auto-detection and debugging support for Django applications via the Python debug adapter.
Auto-detection and debugging support for Flask applications via the Python debug adapter.
Auto-detection and debugging support for Jest test suites via the Node.js debug adapter.
Auto-detection and debugging support for Mocha test suites via the Node.js debug adapter.
Debugging support for Node.js applications via the js-debug adapter.
Debugging support for Python applications via the debugpy adapter.
Observes React component lifecycles, state/prop diffs, and detects anti-patterns during browser recording.
Debugging support for Rust applications via the CodeLLDB adapter.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@KrometrailStart recording browser session on localhost:3000"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Krometrail
Browser memory for coding agents.
A screenshot shows your agent what a page is. Krometrail preserves what the page did—flicker, reversed motion, transient layout shifts, hydration flashes, focus jumps, and incorrect canvas frames that disappear before the next screenshot.
Krometrail controls and records a local Chrome-compatible browser, then turns a selected interval into still-image evidence your agent can inspect: storyboards, difference maps, focused filmstrips, motion history, and the underlying source frames.
Read the documentation · Install · Use it with your agent
Install for Claude Code or Codex
The native plugin supplies the Krometrail skill, MCP connection, and matching managed binary.
# Claude Code
claude plugin marketplace add nklisch/krometrail --scope user
claude plugin install krometrail@krometrail --scope user
# Codex
codex plugin marketplace add nklisch/krometrail
codex plugin add krometrail@krometrailRestart or reload your agent after installation. The first activation downloads and verifies the release paired with the plugin; later starts use the verified local copy.
Related MCP server: markupR MCP Server
Install the standalone command
Use the standalone command for terminal checks or manual MCP configuration on Linux or macOS:
curl -fsSL https://krometrail.dev/install.sh | sh
krometrail --version
krometrail doctorThen register it with an MCP client if needed:
claude mcp add --scope user krometrail -- krometrail mcp
codex mcp add krometrail -- krometrail mcpAsk your agent
Use Krometrail to reproduce the settings-panel animation. It sometimes moves backward before it settles. Inspect the temporal evidence around that interaction, show me the relevant source frames, and tell me whether capture gaps limit the conclusion.
Krometrail preserves evidence; the coding agent still interprets it, finds the responsible code, and verifies the fix.
Local by default
Browser control, recording, captured frames, browser events, and generated artifacts run locally. Krometrail does not send session contents or telemetry to an external service by default. A connected agent reads evidence through local MCP only when it needs it.
Recording is bounded by a configurable disk budget. Krometrail reports known capture gaps and does not claim to capture every browser-rendered frame.
Develop Krometrail
Krometrail is a Rust 2024 workspace with a Rust 1.88 minimum, checked and tested on Linux. The repository selects stable Rust for developer tooling; minimum-compiler compatibility is gated separately. Run the stable quality gate from the repository root:
cargo fmt --all -- --check
cargo check --workspace --all-targets --locked
cargo test --workspace --all-targets --locked
cargo clippy --workspace --all-targets --locked -- -D warningsRun the development binary through Cargo:
cargo run -- --help
cargo run -- doctorSee the development guide for workspace, documentation, release, and fixture commands. The authoritative product and architecture documents remain under docs/.
License
MIT
This server cannot be deployed
Maintenance
Related MCP Connectors
Live browser debugging for AI assistants — DOM, console, network via MCP.
AI agent observability for production traces, natural-language insights, and improvement loops.
Browser-based QA for AI-built software. Test pages with real browsers via agents.
Data + AI observability — monitor and troubleshoot production-grade agents and the context they use.
Related MCP Servers
- AlicenseAqualityBmaintenanceRecords masked browser sessions (rrweb DOM + console + network) to a local SQLite store and exposes them to AI coding agents via stdio MCP. Fully local, no SaaS, no telemetry.107Apache 2.0
- AlicenseNot gradedqualityBmaintenanceEnables AI coding agents to capture screen and voice recordings, extract timestamped frames, and receive structured Markdown reports with context for bug fixing and UI feedback.17 npm18MIT
- AlicenseNot gradedqualityDmaintenanceRecords, replays, and correlates visual and API events in the browser, enabling AI assistants to understand which API calls feed which UI elements.MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI coding agents to visually interact with frontend apps by taking screenshots, clicking elements, reading console logs, and performing visual diffs.20 npm3MIT