Librus MCP Server
An MCP (Model Context Protocol) server that provides AI assistants with access to the Librus Synergia electronic gradebook. It supports multiple student accounts simultaneously and exposes tools for grades (numeric, GPA, and descriptive), messages, attendance, homework, schedules, timetables, announcements, completed lessons, and student information.
Quick Start
You do not need to clone this repository or install Python yourself. The setup uses uv, which installs the correct Python version and Librus MCP automatically.
You need your parent login and password for Librus Synergia. Accounts requiring interactive two-factor authentication are not supported by the underlying library.
1. Install uv
On macOS or Linux, open a terminal and run:
curl -LsSf https://astral.sh/uv/install.sh | sh
On Windows, open PowerShell and run:
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
Close and reopen your AI assistant after installing uv. You can verify the
installation in a new terminal with uvx --version.
2. Create a private credentials file
Create a file named secrets.json in a private folder under your user account,
outside projects and shared or synchronized folders:
{
"accounts": [
{
"alias": "daughter",
"username": "12345",
"password": ""
}
]
}
Replace the empty password value with your Librus password. The alias is the
short name you will use when asking your assistant about this student. For more
than one child, add another account object to the array.
Use the absolute path to this file in the next step. ~ is not expanded in
LIBRUS_CONFIG. Example paths:
- macOS:
/Users/YOUR_NAME/.config/librus-mcp/secrets.json - Linux:
/home/YOUR_NAME/.config/librus-mcp/secrets.json - Windows JSON:
C:\\Users\\YOUR_NAME\\AppData\\Roaming\\librus-mcp\\secrets.json
On macOS or Linux, protect the finished file so only your user can read it:
chmod 600 /absolute/path/to/secrets.json
3. Connect your AI assistant
Choose only the client you use. Replace /absolute/path/to/secrets.json with
the path created above.
Claude Desktop
Open Settings > Developer > Edit Config and add this MCP server. The config
file is normally at ~/Library/Application Support/Claude/claude_desktop_config.json
on macOS or %APPDATA%\Claude\claude_desktop_config.json on Windows.
{
"mcpServers": {
"librus": {
"command": "uvx",
"args": ["librus-mcp==1.2.2"],
"env": {
"LIBRUS_CONFIG": "/absolute/path/to/secrets.json"
}
}
}
}
Claude Code
Run this once in a terminal:
claude mcp add --scope user --transport stdio librus \
-e LIBRUS_CONFIG=/absolute/path/to/secrets.json \
-- uvx librus-mcp==1.2.2
Gemini CLI
Run this once in a terminal. User scope keeps school credentials out of project files.
gemini mcp add --scope user --transport stdio \
-e LIBRUS_CONFIG=/absolute/path/to/secrets.json \
librus uvx librus-mcp==1.2.2
OpenAI Codex CLI
Run this once in a terminal:
codex mcp add librus \
--env LIBRUS_CONFIG=/absolute/path/to/secrets.json \
-- uvx librus-mcp==1.2.2
Never put a Librus password or LIBRUS_ACCOUNTS in a project-level MCP
configuration. Project files can be committed, synchronized, or shared.
4. Restart and verify
- Completely restart Claude Desktop, or reconnect Librus in your CLI client.
- Ask: "Use Librus to list the configured students."
- Check that your chosen aliases appear.
- Ask: "Use Librus to show grades for daughter." Replace
daughterwith your alias. This second request performs a real Librus login.
For CLI status checks, use claude mcp list, gemini mcp list, or
codex mcp list. Claude Code and Gemini CLI also expose status through /mcp.
Troubleshooting
uvxnot found: runuvx --versionin a new terminal, then completely restart the AI assistant. GUI applications may require the absolute path touvx.- Config file does not exist: use an absolute path, not
~. In JSON on Windows, write each backslash twice, for exampleC:\\Users\\.... - Aliases appear but grades fail: verify the same parent credentials at synergia.librus.pl.
- Login requires a second factor: interactive 2FA accounts are not currently supported.
Configuration Reference
Credential sources are checked in this order and are not merged:
LIBRUS_ACCOUNTS, if set.- The absolute file path in
LIBRUS_CONFIG, if set. secrets.jsonin the server working directory.secrets.jsonbeside a source checkout.
Invalid higher-priority configuration produces an error instead of silently
falling back. LIBRUS_FEATURES, LIBRUS_STATE_DIR, and LIBRUS_DOWNLOAD_DIR
provide separate overrides for optional tools and local storage.
For advanced environments, LIBRUS_ACCOUNTS accepts a JSON array directly:
[{"alias":"daughter","username":"12345","password":"..."}]
Alternative: Install from Source
If you prefer to run from a local clone:
git clone https://github.com/krzysztofbury/librus-mcp.git
cd librus-mcp
uv venv && uv pip install -e .
Point LIBRUS_CONFIG at the private credentials file created during Quick Start,
then use the local executable in your MCP config:
{
"mcpServers": {
"librus": {
"command": "/path/to/librus-mcp/.venv/bin/librus-mcp",
"env": {
"LIBRUS_CONFIG": "/absolute/path/to/secrets.json"
}
}
}
}
Contributors can run the credentialed source-tree smoke test with
uv run python verify_connection.py --all-accounts.
Available Tools
| Tool | Description |
|---|---|
list_students() |
List configured student aliases |
get_grades(student_alias, sort_by?) |
Get numeric grades, GPA, and descriptive grades (all, week, or last_login) |
get_final_grades(student_alias) |
Get end-of-year summary per subject: midterm, predicted annual (przewidywana roczna), and annual grade |
get_messages(student_alias, page?, folder?, all_pages?) |
Get one page of messages from the received or sent folder, or the whole folder (bounded at 2000 messages) |
get_message_content(student_alias, message_id) |
Get a message: author, title, date, and content |
get_attendance(student_alias, sort_by?) |
Get attendance records (all, week, or last_login) |
get_attendance_detail(student_alias, detail_url) |
Get details of one attendance entry by its numeric Librus ID |
get_attendance_frequency(student_alias) |
Get attendance frequency per semester and overall |
get_subject_frequency(student_alias, start?, end?) |
Get per-subject attendance percentage, optionally filtered by date range |
get_homework(student_alias, date_from?, date_to?) |
Get homework for a date range (default: next 2 weeks) |
get_homework_detail(student_alias, detail_url) |
Get full details of a homework assignment by its numeric Librus ID |
get_schedule(student_alias, year, month) |
Get calendar events/exams for a month |
get_schedule_detail(student_alias, href) |
Get details of one schedule event (test scope, room, teacher) |
get_recent_schedule_events(student_alias) |
Get schedule events added since the last Librus login. This consumes a read-once Librus view and safely checkpoints its events locally |
get_timetable(student_alias, monday?) |
Get a week's timetable (default: current week; monday picks another week) |
get_announcements(student_alias) |
Get school announcements |
get_completed_lessons(student_alias, date_from, date_to) |
Get completed lessons (subject, teacher, topic) for a date range |
get_student_information(student_alias) |
Get student profile (name, class, tutor, school, lucky number) |
All tools carry MCP ToolAnnotations (read-only / destructive / idempotent
hints), so MCP hosts can apply their own safety policies.
Optional tools (feature gates)
These tools are registered based on the features section of the configuration
(or the LIBRUS_FEATURES env var, a JSON object merged over it):
| Tool | Feature gate | Default | Description |
|---|---|---|---|
get_new_notifications(student_alias) |
notifications |
on | Everything new since the previous call (grades, attendance, messages, announcements, schedule, homework). Seen-state is persisted per student |
get_message_attachments(student_alias, message_id) |
attachments |
on | List attachments (filename + file ID) of a message |
download_attachment(student_alias, message_id, file_id) |
attachments |
on | Download an attachment to the download directory |
get_behaviour_notes(student_alias) |
behaviour_notes |
on | Behaviour notes (uwagi): date, teacher, category, content |
get_recipient_groups(student_alias) |
send_message |
off | List recipient groups for messaging |
get_recipients(student_alias, group) |
send_message |
off | List recipients (name → ID) in a group |
send_message(student_alias, title, content, recipient_ids, confirm_token?) |
send_message |
off | Send a real message to school staff — enable deliberately |
send_message uses a two-step confirmation: the first call sends nothing and
returns a preview plus a single-use confirm_token (valid 5 minutes); only a
second call with that token delivers the message.
Librus exposes recently added schedule events only once. Librus MCP checkpoints each event locally before continuing. After that checkpoint succeeds, a cancelled call or failed state update can replay the event instead of losing it. Recovery uses at-least-once delivery, so a replayed schedule event may appear again. A connection failure, parse failure, local storage failure, or machine crash before the checkpoint completes can still lose data from this read-once upstream view.
Example with all options:
{
"accounts": [{"alias": "daughter", "username": "12345", "password": "..."}],
"features": {
"notifications": true,
"attachments": true,
"behaviour_notes": true,
"send_message": false
},
"state_dir": "~/.librus-mcp/state",
"download_dir": "~/.librus-mcp/downloads"
}
| Setting | Env override | Default | Purpose |
|---|---|---|---|
features |
LIBRUS_FEATURES |
see above | Enable/disable optional tools |
state_dir |
LIBRUS_STATE_DIR |
~/.librus-mcp/state |
Per-student seen-notification state |
download_dir |
LIBRUS_DOWNLOAD_DIR |
~/.librus-mcp/downloads |
Attachment download target |
Attachment files are streamed to an exclusive temporary file and published atomically without overwriting existing files. The download directory must be on a filesystem that supports hard links.
Project Structure
src/
server.py # MCP server with tool definitions and entry point
librus_client.py # Librus API client wrapper with caching and retry
config.py # Environment and JSON-file configuration loader
notification_state.py # Seen IDs and read-once schedule recovery spool
scraping.py # Own Synergia scraping: attachments, behaviour notes
tests/ # pytest suite with mocked librus-apix
Acknowledgments
This project is built on the excellent librus-apix library by RustySnek. Its reverse-engineered Librus client makes this MCP server possible.
Contributing
Contributions are welcome! See CONTRIBUTING.md for guidelines.
Security
To report vulnerabilities, see SECURITY.md.
License
Copyright (C) 2026 Krzysztof Bury.
This project is licensed under the GNU General Public License v3.0 only
(GPL-3.0-only). See LICENSE for the complete terms.
The project adopted GPL-3.0-only in v1.2.2 to honor the strongest license terms
actually distributed with its required librus-apix==1.5.1 dependency. That
dependency declares MIT in package metadata, but both its source repository and
the LICENSE files bundled in its PyPI wheel and source archive contain the
complete GPL-3.0 text. Using GPL-3.0-only avoids relying on the contradictory
metadata when redistributing this combined application.
Metadata
Release files for librus-mcp 1.2.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| librus_mcp-1.2.2.tar.gz | 173.5 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| librus_mcp-1.2.2-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 225.8 kB
Release files / librus_mcp-1.2.2.tar.gz
| Download URL | librus_mcp-1.2.2.tar.gz |
|---|---|
| Size | 173.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a301353ca7c30e88592ab787eb41bebcaeda40f4fa0bc12e72ae1c4194b49caf
|
|
BLAKE2b-256 checksum How to use checksums |
eb7884cb5fb5b96324816a0a7451c9a49acf009adc7b2dbee220900487da4809
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 15, 2026.
Transparency logRelease files / librus_mcp-1.2.2-py3-none-any.whl
| Download URL | librus_mcp-1.2.2-py3-none-any.whl |
|---|---|
| Size | 52.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
cdc49e90e46ceee8e8664c077b4afeff094aca348d43e26d6b98efd46d0b0285
|
|
BLAKE2b-256 checksum How to use checksums |
97b62ef033bba718d4fb5d3a79cdee16c0bce6821564cba907b23f702f6815c0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 15, 2026.
Transparency log