Skip to main content
Glama

DeepLearning.AI MCP Server (dlai-mcp)

License: MIT Python 3.10+ Model Context Protocol

A Model Context Protocol (MCP) server providing AI agents (Claude Desktop, Cursor, Antigravity, custom agent loops) with direct programmatic access to DeepLearning.AI Short Courses on learn.deeplearning.ai.

Designed for autonomous note-taking, code & lab extraction, and instant course research.


Features

  • 📚 Full Course Catalog Access: Browse 100+ DeepLearning.AI short courses with partners, topics, and duration.

  • 📑 Complete Syllabi & Outlines: Query module hierarchies, lesson indices, durations, and lesson types.

  • 🎙️ Timestamped Video Transcripts: Extract caption tracks with structured timestamp markers ([00:14]) for exact citation.

  • 💻 Jupyter Lab & Code Extraction: Read .ipynb notebooks and code exercises rendered in clean Markdown.

  • ⚡ Zero-Auth Direct API: High-speed public tRPC queries without browser overhead.

  • 🔍 Offline SQLite Cache & FTS5 Search: Instant full-text search with BM25 ranking across all synced materials.

  • 🔐 Interactive CLI Login: Simple one-time login command (dlai-mcp login) to authenticate your account.


Related MCP server: Coursera MCP

Installation

Using uv (Recommended)

# Clone the repository
git clone https://github.com/WilliamAxelC/deeplearning.ai-MCP-Server.git
cd deeplearning.ai-MCP-Server

# Create virtual environment and install
uv venv
source .venv/bin/activate
uv pip install -e .

Using pip

pip install -e .

Quickstart & CLI Commands

1. Interactive Authentication (Optional for Public Data, Required for Enrolled Labs)

dlai-mcp login

Opens a Chromium browser window for you to log into your DeepLearning.AI account.

2. Check Status & Local Cache Statistics

dlai-mcp status

3. Pre-fetch Course Transcripts for Offline Agent Access

# Sync a specific course by slug
dlai-mcp sync --course chatgpt-prompt-eng

# Or sync all courses
dlai-mcp sync --all

4. Search Indexed Course Transcripts via CLI

dlai-mcp search "temperature parameter"

5. Start the MCP Server

# Standard stdio mode (for Claude Desktop / Cursor / Antigravity)
dlai-mcp serve

# Or SSE / HTTP network mode
dlai-mcp serve --transport sse --port 8000

Configuring MCP Clients

Claude Desktop Configuration (claude_desktop_config.json)

{
  "mcpServers": {
    "deeplearning-ai": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/deeplearning.ai-MCP-Server",
        "run",
        "dlai-mcp",
        "serve"
      ]
    }
  }
}

Cursor / Antigravity Configuration

{
  "mcpServers": {
    "deeplearning-ai": {
      "command": "/path/to/deeplearning.ai-MCP-Server/.venv/bin/dlai-mcp",
      "args": ["serve"]
    }
  }
}

Available MCP Tools

Tool

Parameters

Description

list_courses

query: Optional[str], partner: Optional[str], topic: Optional[str]

List and filter short courses.

get_course_structure

course_slug: str

Retrieve syllabus, modules, and all lessons.

get_lesson_transcript

course_slug: str, lesson_slug_or_index: str, include_timestamps: bool

Fetch timestamped transcript text.

get_lesson_lab

course_slug: str, lesson_slug_or_index: str, raw_json: bool

Extract Jupyter notebook code & markdown.

search_course_content

query: str, course_slug: Optional[str], limit: int

SQLite FTS5 BM25 search across transcripts.

sync_course

course_slug: str, force_refresh: bool

Pre-fetch and index an entire course into SQLite.

get_auth_status

none

Check active user session and account details.


Author

  • William Axel C (@WilliamAxelC) - williamaxelcuangdinata@gmail.com


License

MIT

Available Tools

7 tools
get_auth_statusA

Check the current DeepLearning.AI authentication status and user account info.

Inputs: no named parameters. Returns: result

get_course_structureA

Retrieve the complete syllabus and lesson outline for a DeepLearning.AI course. Returns module hierarchy, lesson indices, slugs, durations, and lesson types (video, video_notebook, quiz).

Inputs: course_slug (required). Returns: result

get_lesson_labA

Get the Jupyter Lab notebook and code for a lesson. Returns clean interleaved Markdown (explanations + code blocks) by default, or raw .ipynb JSON if raw_json=True.

Inputs: course_slug (required), lesson_slug_or_index (required), raw_json. Returns: result

get_lesson_transcriptA

Get the full timestamped transcript for a lesson in a DeepLearning.AI course. Pass either the lesson index (e.g. '1', '2') or the lesson slug (e.g. 'creating-an-app-with-ai').

Inputs: course_slug (required), lesson_slug_or_index (required), include_timestamps. Returns: result

list_coursesA

List available Short Courses from DeepLearning.AI. Filter by keyword search query, partner organization (e.g. OpenAI, LangChain, Anthropic), or topic.

Inputs: query, topic, partner. Returns: result

search_course_contentA

Full-text search across indexed transcripts, lesson materials, and notebooks using SQLite FTS5 with BM25 ranking. Returns matched snippets, timestamp citations, and direct URLs.

Inputs: query (required), limit, course_slug. Returns: result

sync_courseA

Pre-fetch and index all lesson transcripts and syllabus for a course into the local SQLite database. Enables instant offline search and note-taking.

Inputs: course_slug (required), force_refresh. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.1/5.0

    Scored across 7 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: course listing, course structure, lesson transcript, lesson lab, course-level search, sync/offline indexing, and auth status. There is no meaningful overlap or ambiguity among the tool names or their responsibilities.

    Naming Consistency5/5

    Tools consistently use snake_case verb_noun naming: list_courses, get_ourse_structure, get_lesson_transcript, search_course_ontent, sync_course. The pattern is obvious and predictable for an agent routing to tools.

    Tool Count5/5

    Seven tools fit the server's scope well: discovery, course inspection, lesson content retrieval, search, and sync. It feels complete without unnecessary bulk or missing core steps.

    Completeness4/5

    The main learning workflow is covered: find a course, get its structure, fetch lesson transcripts and labs, and search indexed content. Minor gaps exist, such as no direct quiz content fetch and only a limited auth status check, but these do not block the core set.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      Enables AI agents to browse and teach the Deep Learning DIY course, providing access to modules, notebooks, transcripts, quizzes, slides, and homework via MCP tools.
      18
      4
      Apache 2.0
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI agents to programmatically interact with Coursera, including enrolling in courses, completing lectures, solving quizzes, submitting assignments, earning certificates, and pushing them to LinkedIn.
      1
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      Enables fetching enrolled course lists, lessons, transcripts, video frames, and slides from DeepLearning.AI short courses, so an AI can summarize them and pass notes to other tools.
      6
      MIT
    • A
      license
      A
      quality
      B
      maintenance
      Enables AI agents to search 1m+ open-access papers, standards, textbooks, and legal documents, resolve identifiers such as RFCs, DOIs, and legal citations, download PDFs, and retrieve citations and metadata through MCP tools.
      9
      MIT