What it does
Diffgrab detects changes on tracked URLs by hashing page content. Each diff combines a unified diff with section-level analysis that shows which headings changed, plus a short readable summary of lines added and removed per section. Page content is extracted with markgrab, which handles HTML, YouTube, PDF and DOCX sources. An optional visual diff compares screenshots taken with snapgrab.
It suits anyone who needs to know when a web page changes and what changed, for example from a Claude Code session. Its five MCP tools register a URL, check tracked URLs, return a diff between snapshots, browse snapshot history and stop tracking a URL. A command-line interface offers matching track, check, diff, history and untrack commands, and the Python API returns a DiffResult for every diff operation.
Snapshots are saved in a SQLite database that is created automatically in the user's local data directory, and a custom path can be set. Running it means adding the server to a Claude Code MCP configuration or launching it with uvx. Optional extras are installed for the visual diff feature. The project is released under the MIT license.
Limitations and requirements
Visual screenshot comparison depends on the separate snapgrab package, which is optional and must be installed through the extras.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
Diffgrab
Web page change tracking with structured diffs. markgrab + snapgrab integration.
v0.1.0As shown on registry.modelcontextprotocol.io when we last read it