About wallabag MCP
wallabag MCP provides wallabag users and AI assistants with tools for working with articles, tags, and annotations. It wraps the OAuth API as a structured interface for read-it-later libraries, including self-hosted instances.
Key features
- 01List and filter paginated entries
- 02Fetch entry metadata and content
- 03Save, update, and archive entries
- 04Manage tags and annotations
- 05Check connectivity and authentication
Use cases
- 01Find and read saved articles
- 02Add a webpage to wallabag
- 03Organize tags and read status
- 04Annotate passages in articles
MCP support
The MCP server exposes wallabag entry, tag, and annotation operations to clients, connecting to an instance with configured OAuth credentials.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
uvx wallabag-mcpWALLABAG_BASE_URLBase URL of the wallabag instance, e.g. https://wallabag.example.comWALLABAG_ACCESS_TOKENSecretExisting OAuth access token; skips password grant if providedWALLABAG_CLIENT_IDOAuth client id; required unless WALLABAG_ACCESS_TOKEN is setWALLABAG_CLIENT_SECRETSecretOAuth client secret; required unless WALLABAG_ACCESS_TOKEN is setWALLABAG_USERNAMEwallabag username; required unless WALLABAG_ACCESS_TOKEN is setWALLABAG_PASSWORDSecretwallabag password; required unless WALLABAG_ACCESS_TOKEN is setLocal servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEwallabag MCP
A Model Context Protocol (MCP) server for wallabag, the open-source read-it-later service.
This server gives AI assistants a structured interface for common wallabag workflows:
- Verify connectivity and authentication with a tiny read-only health check
- List, filter, and page through saved articles
- Fetch entry metadata or full extracted content
- Save new URLs
- Archive/unarchive and star/unstar entries
- Reload/refetch article extraction
- Delete entries
- List and add tags
- List, create, and delete annotations
Why this exists
wallabag is widely used in self-hosted setups, but its API is OAuth-based and awkward for general LLM tools. This MCP server wraps the important entry, tag, and annotation operations with an AI-friendly surface and documented parameters.
Installation
pipx install git+https://github.com/rusty4444/wallabag-mcp.git
Or from a checkout:
python -m venv .venv
source .venv/bin/activate
pip install -e .
Configuration
The server reads configuration from environment variables:
| Variable | Required | Description |
|---|---|---|
WALLABAG_BASE_URL | Yes | Base URL of your wallabag instance, e.g. https://app.wallabag.it |
WALLABAG_ACCESS_TOKEN | Optional | Existing OAuth access token; if set, password grant is skipped |
WALLABAG_CLIENT_ID | Required unless access token is set | OAuth client id from wallabag's developer client page |
WALLABAG_CLIENT_SECRET | Required unless access token is set | OAuth client secret |
WALLABAG_USERNAME | Required unless access token is set | wallabag username |
WALLABAG_PASSWORD | Required unless access token is set | wallabag password |
WALLABAG_TIMEOUT | No | HTTP timeout in seconds, default 20 |
Create a wallabag API client from your wallabag instance under Developer / API clients. wallabag's official docs show the password grant against /oauth/v2/token.
MCP client config
{
"mcpServers": {
"wallabag": {
"command": "wallabag-mcp",
"env": {
"WALLABAG_BASE_URL": "https://wallabag.example.com",
"WALLABAG_CLIENT_ID": "your-client-id",
"WALLABAG_CLIENT_SECRET": "your-client-secret",
"WALLABAG_USERNAME": "your-username",
"WALLABAG_PASSWORD": "your-password"
}
}
}
}
Tools
| Tool | Purpose |
|---|---|
wallabag_health_check | Verify connectivity and authentication with a tiny read-only request |
wallabag_list_entries | List entries with pagination and filters |
wallabag_get_entry | Fetch one entry, optionally including extracted content |
wallabag_add_entry | Save a URL into wallabag |
wallabag_update_entry | Update title, URL, archive/starred state, or tags |
wallabag_archive_entry | Mark an entry archived/read |
wallabag_unarchive_entry | Mark an entry unarchived/unread |
wallabag_star_entry | Star/favourite an entry |
wallabag_unstar_entry | Remove starred/favourite state |
wallabag_reload_entry | Ask wallabag to refetch/reparse the original URL |
wallabag_delete_entry | Delete an entry |
wallabag_list_tags | List known tags |
wallabag_add_tags_to_entry | Add comma-separated tags to an entry |
wallabag_delete_tag | Delete a tag globally |
wallabag_list_annotations | List annotations for an entry |
wallabag_create_annotation | Create an annotation on quoted article text |
wallabag_delete_annotation | Delete an annotation |
Development and validation
python -m venv .venv
source .venv/bin/activate
pip install -e '.[dev]'
ruff check .
pytest
python scripts/live_docs_test.py
live_docs_test.py validates wallabag's public API documentation and the hosted API docs page without needing account credentials. Mutation and authenticated request behaviours are covered with mocked HTTP tests.
Safety
Write-capable tools mutate your read-it-later library. Keep OAuth secrets in environment variables or a secret manager, never in source control.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.