Skip to main content
Glama
akarazhev

PageVault MCP Server

by akarazhev
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Go
Local

PageVault

PageVault saves loaded web page content as structured snapshots in a local SQLite archive. Enable recording for a site and browse normally; changes produce snapshots and the archive retains unique observed blocks. A generic extractor handles common content, and a GitLab adapter adds project metadata. Search, view versions, export JSON, or read the archive through a local MCP client.

PageVault is free and distributed through GitHub Releases. The browser extension and native helper run locally; there is no separate server or telemetry. Saved content can contain private data. Read the privacy policy before enabling a site.

Install and use

Download PageVault-1.4.0.zip from Releases. It includes the extension, installers, six host binaries, source and documentation. The automatically generated GitHub source archives do not contain binaries. On macOS 13 or newer, PageVault-1.4.0-macOS-unsigned.pkg installs the helper for Intel and Apple Silicon; use it with PageVault-Extension-1.4.0.zip or the full bundle. Verify each download against its .sha256 file. The PKG is unsigned and not notarized; macOS may block it. See installation for Apple's individual-file exception and restrictions.

  1. Extract the full bundle to a permanent location. On macOS/Linux run ./install.sh chrome; on Windows run install-windows.cmd or ./install-windows.ps1 -Browser chrome. Choose chromium, edge, or brave as appropriate, and repeat for each browser. The installer displays the helper and database paths.

  2. Open the browser's extensions page, enable developer mode, and load the unpacked extension folder (or the root of the standalone extension ZIP). Keep the folder in place. The fixed extension ID is ijkioljfdaaplhaldbdpabckkfcieiee.

  3. Open PageVault settings and check the helper connection and SQLite path. Visit an HTTP(S) page, enable recording for its site in the extension popup, and grant the requested site permission. Use Capture now for a manual snapshot. Disable recording at any time; saved data stays in SQLite.

See installation, upgrading and backups, and queue recovery. Wait for the queue to empty before removing the extension: its IndexedDB queue belongs to the browser.

Related MCP server: shop-sql-mcp

What is captured

PageVault extracts loaded DOM content including text, headings, lists, tables, code, links and media references. Open shadow roots and accessible same-origin frames are handled where possible. Scroll or expand content to load it first. Closed shadow roots, cross-origin frames, image text, canvas and PDF viewers may be missing; media files are not downloaded. The archive does not recreate an interactive site offline. Snapshot warnings report limits reached. Defaults include 5,000 blocks, 1,000,000 text characters and 2,000,000 HTML characters. Optional sanitized HTML omits scripts, event handlers and form values and is never executed by the archive UI.

The queue accepts up to 500 captures or 128 MiB; new captures are rejected when full, without deleting accepted entries. Restore the helper, retry delivery, then capture again. Permanently rejected entries can be exported, retried or discarded in settings. Accumulated blocks can include multiple revisions; choose a snapshot to see one state.

MCP and development

An MCP client launches pagevault-host mcp --db /absolute/path/archive.sqlite3 over stdio, opening an existing schema 1 database read-only. Tools list sites, search pages, and retrieve accumulated content or snapshot versions. An agent using a cloud model may send retrieved content to its provider. See MCP setup and limits.

Development requires Go 1.27.3, Node 26.12.1, npm 12.2.1 and Playwright Chromium. Run npm ci, go test ./..., npm test, npm run lint, npm run format:check, and npm run check. Build release bundles with go run ./cmd/release, go run ./cmd/release --extension-only, and on macOS go run ./cmd/package-macos --unsigned. See verification and contributing.

Documentation: architecture, schema, native protocol, adapters, distribution, readiness and support. Report security issues privately via GitHub. PageVault is MIT licensed; dependency notices are in THIRD_PARTY_NOTICES.md.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • mcpserve-py

    devaloi
    A
    license
    A
    quality
    D
    maintenance
    Exposes SQLite database query tools and markdown document resources over JSON-RPC 2.0 stdio transport, enabling AI assistants to read and search documents and execute read-only SQL queries.
    8
    1
    MIT

  • shop-sql-mcp

    lampmaster
    F
    license
    Not graded
    quality
    C
    maintenance
    Provides AI agents read-only analytical access to a SQLite database over stdio, with tools for listing tables, describing schemas, and running paginated SQL queries.
    -

  • shop-mcp

    AndrewKonst
    F
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to read-only analyze a SQLite e-commerce database, exploring schema and running analytical SQL queries over stdio.
    -

  • Universal Safe Database MCP

    Shreyansh-awasthi
    A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI models to securely analyze SQLite datasets and read local text documents through a local stdio interface. It provides read-only querying, schema discovery, pagination, and timeout protections to prevent data leaks or destructive changes.
    MIT