tue-api-wrapper MCP server
Commit Nov 12, 2026Provides an interface for AI clients to interact with University of Tübingen study systems including Alma, ILIAS, Moodle, and campus data.
Also known as Tue-StudyOS/tue-api-wrapper
- Stars
- 14
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 12, 2026
- License
- Apache-2.0
What is tue-api-wrapper MCP?
tue-api-wrapper provides an interface for AI clients to interact with University of Tübingen study systems including Alma, ILIAS, Moodle, and campus data.
Summary generated from the project's README
About tue-api-wrapper MCP server
Students, instructors, and staff at the University of Tübingen can use this tool to interact with various campus systems through an AI client. It provides a unified interface for accessing academic data from platforms like Alma, ILIAS, Moodle, TIMMS, and university mail. The server can be run locally using the tue-mcp command, supporting stdio for local agents or streamable-http for HTTP clients.
Accessing public information such as canteen menus, campus events, or course catalogs requires no credentials. However, accessing private student data like personal timetables, Moodle deadlines, or ILIAS tasks requires authentication. Users must configure their university username and password within their local process or via environment variables to ensure sensitive credentials never pass through hosted services. This setup allows for secure, local management of private academic records and communications.
What you can do with it
- 1Plan academic semestersAn instructor can use the tool to coordinate teaching and manage course materials across different university platforms.
- 2Check campus amenitiesA student can ask an AI to find out what is being served in the university canteens on a specific date.
- 3Manage course deadlinesA user can request a summary of upcoming academic deadlines from Moodle for the next month.
- 4Search for lecture materialsA researcher can search for specific lecture recordings or streams within the TIMMS system.
Written from the project's README. Every use case is checked against a quote from it.
Configuration
UNI_PASSWORDUNI_USERNAME
Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachabletue-studyos/tue-api-wrapper is public and reachablegithub.com/tue-studyos/tue-api-wrapper ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-12github.com/tue-studyos/tue-api-wrapper ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubtue-api-wrapper
Python SDK, local API server, MCP tools, and app clients for University of Tuebingen study systems.
The project wraps live university systems such as Alma, ILIAS, Moodle, university mail, TIMMS, campus pages, and public course data. It does not ship mock data or replace those upstream systems. Public features work without credentials; private student features require credentials in your local process.
Install
Install the published Python package:
pip install tue-api-wrapper
Optional extras:
pip install "tue-api-wrapper[mcp]"
pip install "tue-api-wrapper[discovery]"
Install the university-workflows skill
This repository bundles a provider-agnostic agent skill for University of Tuebingen students, instructors, professors, and staff. It supports semester and course planning, teaching coordination, course and material discovery, private university data, documents, registrations, and daily workload planning. Retrieve only the skill directory, not a full repository checkout, then install it into the skill directory specified by your agent provider. For example:
git clone --depth=1 --filter=blob:none --sparse \
https://github.com/Tue-StudyOS/tue-api-wrapper.git /tmp/tuebingen-university-workflows-skill
git -C /tmp/tuebingen-university-workflows-skill sparse-checkout set \
skills/tuebingen-university-workflows
python3 /tmp/tuebingen-university-workflows-skill/skills/tuebingen-university-workflows/scripts/install_skill.py \
--destination /path/to/your-provider-skills
Paste this into an agent to install and begin using it right away:
Install the
tuebingen-university-workflowsagent skill from theTue-StudyOS/tue-api-wrapperrepository into this provider's skills directory. Do not clone the full repository: retrieve onlyskills/tuebingen-university-workflows/with a sparse checkout or equivalent. Determine the correct destination from the provider documentation, then run that directory'sscripts/install_skill.py --destination <provider-skills-directory>, validate the installation, and start the skill's welcome menu rather than only saying it is installed. Do not ask for or handle credentials until I choose a private task. For every registration, enrolment, waitlist, booking, calendar write, or external form submission, inspect first and obtain my final explicit confirmation immediately before submitting.
The skill uses public data without credentials and asks the account owner to
configure private credentials only in their local process. It never asks users
to paste passwords into chat. Re-run the installer with --force only to
replace an existing copy of this same skill.
Python Module Usage
Use TuebingenPublicClient for public data. It does not need a university login.
from tue_api_wrapper import TuebingenPublicClient
client = TuebingenPublicClient()
modules = client.alma.search_modules("machine learning", max_results=10)
lectures = client.alma.current_lectures(date="02.05.2026", limit=20)
canteens = client.campus.canteens()
events = client.campus.events(query="KI", limit=10)
recordings = client.timms.search("theoretische informatik", limit=5)
Use TuebingenAuthenticatedClient for private university data. Credentials stay local to the Python process.
import os
from tue_api_wrapper import TuebingenAuthenticatedClient
client = TuebingenAuthenticatedClient.login(
username=os.environ["UNI_USERNAME"],
password=os.environ["UNI_PASSWORD"],
)
profile = client.alma.profile()
timetable = client.alma.timetable("Sommer 2026")
documents = client.alma.studyservice_documents()
tasks = client.ilias.tasks()
deadlines = client.moodle.deadlines(days=30)
inbox = client.mail.inbox(limit=5)
client.close()
Credentials can also be loaded from .env:
UNI_USERNAME=your-zdv-id
UNI_PASSWORD=your-password
from tue_api_wrapper import TuebingenAuthenticatedClient, UniversityCredentials
credentials = UniversityCredentials.from_env(".env")
client = TuebingenAuthenticatedClient(credentials)
PPI exam protocols
PPI uses an independent account. The ZDV username identifies the student mailbox, but the password must be a separately chosen PPI password.
from tue_api_wrapper import PpiClient
ppi = PpiClient()
signup = ppi.signup("zxabc12", "a-separate-ppi-password")
# Follow the activation link sent to signup.activation_email.
ppi.login("zxabc12", "a-separate-ppi-password")
catalog = ppi.fetch_lecture_catalog()
borrowed = ppi.fetch_borrowed_lectures()
ppi.close()
Common Methods
Public Alma
client.alma.search_modules("informatics", max_results=20)
client.alma.module_search_filters()
client.alma.module_detail(detail_url)
client.alma.current_lectures(date="02.05.2026", limit=20)
Public Campus
client.campus.canteens(menu_date="2026-05-07")
client.campus.canteen(611, menu_date="2026-05-07")
client.campus.buildings()
client.campus.building_detail("/einrichtungen/...")
client.campus.events(query="lecture", limit=24)
client.campus.gym_occupancy()
client.campus.kuf_occupancy()
client.campus.seat_availability()
Public Directory, TIMMS, Talks, and Career Data
client.directory.search("informatik")
client.timms.suggest("analysis", limit=8)
client.timms.search("theoretische informatik", offset=0, limit=20)
client.timms.item("item-id")
client.timms.streams("item-id")
client.timms.tree(node_id="root")
client.talks.search(query="AI", limit=16)
client.talks.item(123)
client.praxisportal.filters()
client.praxisportal.search(query="internship", per_page=20)
client.praxisportal.project(12345)
Authenticated Alma
client.alma.profile()
client.alma.timetable("Sommer 2026")
client.alma.timetable_controls()
client.alma.timetable_view(term="Sommer 2026", limit=50)
client.alma.refresh_timetable_export_url(term="Sommer 2026")
client.alma.timetable_course_assignments("Sommer 2026", limit=20)
client.alma.current_lectures(date="02.05.2026", limit=50)
client.alma.course_offerings(query="data science", term="Sommer 2026")
client.alma.course_registration_support(detail_url)
client.alma.course_registration_options(detail_url)
client.alma.register_for_course(detail_url, planelement_id="...")
client.alma.catalog_page(term="Sommer 2026", limit=80)
client.alma.study_planner()
client.alma.portal_messages()
client.alma.exams()
client.alma.exam_reports()
client.alma.download_exam_report()
client.alma.enrollments()
client.alma.studyservice_documents()
client.alma.download_document(doc_id)
Authenticated ILIAS
client.ilias.root()
client.ilias.memberships()
client.ilias.tasks()
client.ilias.content(target)
client.ilias.forum_topics(target)
client.ilias.exercise_assignments(target)
client.ilias.search_filters()
client.ilias.search("algorithms", page=1)
client.ilias.info(target)
client.ilias.add_favorite(url)
client.ilias.waitlist_support(url)
client.ilias.join_waitlist(url, accept_agreement=True)
Authenticated Moodle
client.moodle.dashboard(event_limit=6, course_limit=12, recent_limit=9)
client.moodle.deadlines(days=30, limit=50)
client.moodle.courses(classification="all", limit=24, offset=0)
client.moodle.categories()
client.moodle.course(course_id)
client.moodle.enrol_in_course(course_id, enrolment_key=None)
client.moodle.grades()
client.moodle.messages()
client.moodle.notifications()
Authenticated Mail and Portal
client.mail.inbox(limit=12)
client.mail.mailbox(name="INBOX", limit=12, unread_only=False, query="")
client.mail.mailboxes()
client.mail.message(uid, mailbox="INBOX")
client.portal.dashboard(term="Sommer 2026", limit=8)
client.portal.search("seminar", term="Sommer 2026")
client.portal.item(item_id, term="Sommer 2026")
client.portal.course_detail(title="Machine Learning", term="Sommer 2026")
Course Discovery
client.discovery.search("machine learning", limit=20)
client.discovery.search(
"seminar",
sources=("alma", "ilias"),
term="Sommer 2026",
include_private=True,
)
client.discovery.refresh(include_private=True, limit=3000)
client.discovery.status()
Local API Server
Start the FastAPI server:
tue-api-server
Useful URLs:
- API root:
http://127.0.0.1:8000/ - health check:
http://127.0.0.1:8000/api/health - OpenAPI docs:
http://127.0.0.1:8000/docs
Public routes work without credentials. Authenticated university routes read
UNI_USERNAME and UNI_PASSWORD. PPI routes instead require the independent
PPI_USERNAME and PPI_PASSWORD values. The PPI surface includes signup,
lecture listing, borrowing, downloads, and token requests below /api/ppi/.
Feedback Issue Creation
The desktop, web, and iOS clients can create public GitHub feedback issues directly from the client without the local API server. Configure a fine-grained GitHub token with access to create issues in this repository:
- Desktop renderer builds read
VITE_GITHUB_FEEDBACK_TOKEN. - Next.js browser builds read
NEXT_PUBLIC_GITHUB_FEEDBACK_TOKEN. - iOS reads the
GITHUB_FEEDBACK_TOKENbuild setting intoGitHubFeedbackTokeninInfo.plist.
If the token is missing, the feedback UI shows that issue creation is unavailable.
Local MCP Server
Install the MCP extra and start the server:
pip install "tue-api-wrapper[mcp]"
tue-mcp
Use stdio for local agent clients. For HTTP clients:
tue-mcp --transport streamable-http --host 127.0.0.1 --port 8765
Repository Layout
| Path | Purpose |
|---|---|
package/ |
Python clients, parsers, SDK, FastAPI routes, MCP server, and tests |
java/ |
Native Java/Android clients for upstream university systems |
nextjs/ |
Next.js dashboard |
desktop/ |
Electron desktop app with local credential handling |
ios/ |
SwiftUI app and native client work |
chatgpt/ |
ChatGPT app, MCP integration, and widget UI |
go/ |
Go CLI experiments for stable request flows |
cli/ |
Repo-local wrapper scripts |
docs/ |
Discovery notes, release notes, SDK docs, and screenshots |
examples/ |
Small example projects and usage snippets |
Development
Run the checks that match your change:
cd package && pytest
cd java && gradle build
npm --prefix desktop run build
npm --prefix nextjs run build
npm run generate:ios
npm run build:ios
When adding integrations, start with the Python package, parse upstream responses into typed contracts, add focused tests, and expose shared JSON through FastAPI when app surfaces need it.
Security
- Do not commit credentials, cookies, HAR files, signed URLs, PDFs, mailbox exports, or session artifacts.
- Keep authenticated flows local to the client or local backend.
- Do not route student credentials through hosted services.
- Prefer public data for teaching and demos unless private account data is required.
- Return clear errors for missing credentials or upstream failures.
License
This repository is licensed under the Apache License 2.0. See LICENSE.
tue-api-wrapper MCP questions
Does the server provide mock data?
No, the project wraps live university systems and does not ship mock data or replace upstream systems.
How do I access private student features?
Private features require credentials provided in your local process; public features work without credentials.
Is it safe to paste my password into the chat?
No, the system is designed to keep credentials local and never asks users to paste passwords into chat.