Onboarding MCP Server
Enables writing structured Markdown notes and Maps of Concepts (MOCs) directly into a local Obsidian vault.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Onboarding MCP Server Map the codebase in ~/code/myapp and write a summary note to Obsidian"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Onboarding MCP Server 🚀
A local-first Model Context Protocol (MCP) server that empowers AI coding assistants (e.g., Cursor, Claude Code, Antigravity IDE) to autonomously map codebases, transcribe onboarding videos with keyframe screenshots, and maintain structured knowledge vaults in Obsidian directly from IDE chat interactions.
🌟 Key Capabilities
set_vault_path: Dynamically configure the active Obsidian vault directory for session state from IDE chat.set_workspace_path: Dynamically configure the target codebase workspace directory for session state from IDE chat.get_codebase_structure: Recursively map code directory structures and detect dependency manifests, respecting.gitignore(uses configured session workspace path or optionalworkspace_pathparameter).process_video: Transcribe onboarding videos usingfaster-whisper(CPU int8), extract scene keyframes via OpenCV, and output timestamp-aligned transcript data.write_obsidian_note: Write structured Markdown notes and MOCs (Maps of Concepts) directly into the user's local Obsidian vault (uses configured session vault path or optionalvault_pathparameter).
Related MCP server: knitbrain
🚀 Quickstart
Option 1: Docker (Recommended - Zero Host Setup)
Build and run the containerized server with pre-packaged ffmpeg and dependencies. Note: because Docker containers run in an isolated environment, volume mounts (-v) are required to grant the container access to host directories specified in chat:
# Build the Docker image
make build-docker
# Run container (mount host directories so chat can access them)
docker run -i --rm -v $HOME:$HOME onboarding-mcp-serverOption 2: Local Python Environment (No Volume Mounts Needed)
Runs directly on your host OS. Any absolute path passed from IDE chat works natively without volume mounts.
Requires Python 3.11+ and ffmpeg installed on host (brew install ffmpeg on macOS).
# Install dependencies into virtualenv
make setup
# Run the tests
make test
# Run the MCP server over stdio
.venv/bin/python -m onboarder.server⚙️ IDE Chat Integration & Setup
The IDE chat assistant can configure workspace and vault paths dynamically at any point during a chat session:
set_workspace_path(path="/path/to/repo")set_vault_path(path="/path/to/ObsidianVault")
Alternatively, individual tools (get_codebase_structure, write_obsidian_note, process_video) accept optional workspace_path and vault_path arguments directly.
Docker Configuration Example (claude_desktop_config.json / Cursor / Antigravity)
Mount your user home directory (or relevant code/vault folders) so paths passed in chat are accessible:
{
"mcpServers": {
"onboarding-mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-v",
"/Users/yourname:/Users/yourname",
"onboarding-mcp-server"
]
}
}
}Local Python Configuration Example
{
"mcpServers": {
"onboarding-mcp": {
"command": "/absolute/path/to/onboarding-mcp/.venv/bin/python",
"args": ["-m", "onboarder.server"]
}
}
}🛠 Developer Workflow
make setup # Set up venv and install dev dependencies
make test # Run pytest suite
make lint # Run ruff check and mypy type checking
make format # Run ruff formatter📄 License
MIT
Available Tools
5 toolsget_codebase_structureA
Recursively scan workspace directory and return file tree and dependencies, respecting .gitignore.
Inputs: workspace_path. Returns: result
process_videoA
Process onboarding video: extract audio, transcribe with Whisper, extract scene keyframes, and return timestamp-aligned JSON payload.
Inputs: video_path (required), output_dir, vault_path, extract_frames. Returns: result
set_vault_pathB
Set the active Obsidian vault directory path for the session state.
Inputs: path (required). Returns: result
set_workspace_pathA
Set the active workspace directory path for the session state.
Inputs: path (required). Returns: result
write_obsidian_noteC
Write a markdown note into the active Obsidian vault directory.
Inputs: content (required), note_path (required), vault_path. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 5 tools
Tools are mostly distinct, but set_vault_path and set_workspace_path have near-identical names and purposes, differing only by target directory. The other three tools (get_codebase_structure, process_video, write_obsidian_note) are clearly distinct actions.
All tool names follow a consistent verb_noun pattern: set, set, get, process, write. The nouns are descriptive and uniform in style, making the API predictable.
Five tools are well-scoped for an onboarding workflow, covering path setup, codebase analysis, video processing, and note writing without unnecessary bloat.
The toolset covers the core onboarding steps, but there are minor gaps such as no tool to list or manage existing notes, and no way to inspect the video metadata before processing. These are workable but not exhaustive.
Maintenance
Related MCP Connectors
An MCP server that gives your AI access to the source code and docs of all public github repos
Markdown-based note-taking with a hosted MCP server. Your notes serve you and your AI.
Driflyte MCP server which lets AI assistants query topic-specific knowledge from web and GitHub.
Shared memory MCP server for Claude Code, Codex, Gemini CLI and Cursor. Cited, editable.
Related MCP Servers
- AlicenseAqualityBmaintenanceA local-first MCP server that gives AI coding agents persistent memory and controlled commands. Features a git-backed markdown knowledge vault with FTS5 search, surgical section edits, token-aware context budgeting, and a sandboxed command engine with human approval gates. Works with Claude Code, Cursor, Copilot, Gemini, and more.621Apache 2.0
- AlicenseAqualityCmaintenanceLocal-first MCP server that gives any AI coding agent per-project memory, workflow intelligence, and always-on, lossless token & context optimization.3731 npm6MIT
- FlicenseNot gradedqualityDmaintenanceA local MCP server that gives AI agents structured access to a personal Obsidian knowledge vault, with semantic search, organization through Maps of Content, and git-backed history.-

Vibgrate AI Contextofficial
AlicenseNot gradedqualityBmaintenanceLocal-first MCP server that gives AI assistants codebase intelligence—code graph, drift analysis, vulnerability attribution, and version-correct library docs—all from the user's machine.809 npm6Apache 2.0