loom-mcp
Loom MCP server exposes many tools to read and manage Loom videos and workspace data through MCP clients.
Discover videos: list, search, get metadata/details, counts, watch-later status, and last watch time.
Read content: transcripts, WebVTT captions, AI summaries, descriptions, chapters, key takeaways, tasks, comments, reactions, tags, backlinks, meeting notes, Confluence pages, user profiles, and workspace tags.
Access media: get signed MP4 download URLs, regenerate MP4s, and per README download audio/video with optional trimming.
Save to disk: most per-video read tools accept
save_dirand write files likemetadata.json,transcript.txt,captions.vtt,summary.txt, and combineddetails.md.Manage data: rename/update videos and settings; create, edit, delete, approve, or respond to comments/tasks; add/delete reactions; follow/unfollow videos and tags; archive, delete, recover, duplicate, pin, or move videos; manage folders; add/remove watch later; share videos to spaces.
Integrate: works with Claude, Cursor, VS Code, etc., using a Loom browser session cookie (
connect.sid) and Python 3.11+/uvx.
Allows fetching linked Confluence pages and meeting notes URLs from Loom videos.
Provides tools to list, search, and manage Loom videos, including transcripts, summaries, comments, tasks, reactions, tags, folders, spaces, and more via Loom's internal GraphQL API.
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., "@loom-mcp list my recent videos"
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.
Loom MCP Server
MCP server exposing 59 tools for Loom's internal GraphQL API. Works with Claude, Cursor, or any MCP-compatible client.
Features
List, search, and get detailed metadata for Loom videos
Read transcripts, captions, AI summaries, chapters, and key takeaways
Save any fetched content to disk on demand via
save_dirparameterManage comments, tasks, reactions, and tags
Organize with folders, spaces, and watch lists
Update video settings, share to spaces, and more
Related MCP server: mcp-video-analyzer
Prerequisites
Python 3.11+
uv package manager
Setup
Authentication
This server uses Loom's internal GraphQL API via a browser session cookie. There is no official API key — you must grab the cookie from your browser.
Open loom.com in your browser
Open DevTools (F12) → Application → Cookies →
https://www.loom.comCopy the value of the
connect.sidcookie (starts withs%3A...)Paste it as
LOOM_COOKIEin your MCP client config below, prefixed withconnect.sid=
The cookie lasts about 30 days. See Troubleshooting if you get auth errors.
Installation
Pick your MCP client below for the appropriate config. Each uses uvx to install and run directly from GitHub — no clone needed.
Add to your claude_desktop_config.json:
{
"mcpServers": {
"loom": {
"type": "stdio",
"command": "uvx",
"args": ["--from", "git+https://github.com/karbassi/loom-mcp.git", "loom-mcp"],
"env": {
"LOOM_COOKIE": "connect.sid=s%3A..."
}
}
}
}claude mcp add loom -- uvx --from git+https://github.com/karbassi/loom-mcp.git loom-mcpThen set the env var in your shell or .env:
export LOOM_COOKIE="connect.sid=s%3A..."Add to your Cursor MCP settings (.cursor/mcp.json):
{
"mcpServers": {
"loom": {
"type": "stdio",
"command": "uvx",
"args": ["--from", "git+https://github.com/karbassi/loom-mcp.git", "loom-mcp"],
"env": {
"LOOM_COOKIE": "connect.sid=s%3A..."
}
}
}
}Add to your VS Code settings (.vscode/mcp.json):
{
"mcpServers": {
"loom": {
"type": "stdio",
"command": "uvx",
"args": ["--from", "git+https://github.com/karbassi/loom-mcp.git", "loom-mcp"],
"env": {
"LOOM_COOKIE": "connect.sid=s%3A..."
}
}
}
}git clone git@github.com:karbassi/loom-mcp.git
cd loom-mcp
uv sync
uv run loom-mcpOr reference it from any MCP client:
{
"mcpServers": {
"loom": {
"type": "stdio",
"command": "uv",
"args": ["run", "--directory", "/path/to/loom-mcp", "loom-mcp"],
"env": {
"LOOM_COOKIE": "connect.sid=s%3A..."
}
}
}
}Shared HTTP server
By default every client starts its own stdio process. To run one long-lived server that many clients share:
LOOM_COOKIE="connect.sid=s%3A..." LOOM_MCP_TRANSPORT=http LOOM_MCP_PORT=8505 loom-mcp{
"mcpServers": {
"loom": { "type": "http", "url": "http://127.0.0.1:8505/mcp" }
}
}The endpoint has no authentication: anything that can connect to it can use every tool with your Loom session, including deleting videos. So:
The server only binds to loopback (
LOOM_MCP_HOSTmust be127.0.0.1,localhost, or::1). Any local process can still connect, so only use HTTP mode on a machine where every local user and process is trusted with your Loom account.It always turns on FastMCP's Host/Origin protection, which FastMCP leaves off by default. That check only stops DNS-rebinding attacks from web pages. It does not authenticate anyone.
To reach it from another machine, put it behind a reverse proxy that controls who can connect: a bearer token, mTLS, an IP allowlist on a private network, or similar. Then add the proxy's hostname to
FASTMCP_HTTP_ALLOWED_HOSTS, e.g.FASTMCP_HTTP_ALLOWED_HOSTS='["mcp.example.com"]'. If a browser-based client will call it, also add that client's origin toFASTMCP_HTTP_ALLOWED_ORIGINS, e.g.'["https://mcp.example.com"]'. Never expose the port or the proxy publicly without access control.
File-writing tools (save_dir on the read tools, download_media) write on the server's machine, not the client's, and need absolute paths. See Saving to disk.
Tools
Read (31 tools)
Tools marked with Save accept an optional save_dir parameter — see Saving to disk.
Tool | Description | Save |
| List your videos, sorted by most recent | |
| AI-powered semantic search | |
| Video metadata (name, duration, owner, views) |
|
| Full transcript with timestamps and speakers |
|
| WebVTT captions with start+end timestamps per cue |
|
| AI-generated summary |
|
| AI-generated chapters |
|
| AI-generated detailed description with timestamped sections |
|
| AI-generated key takeaways |
|
| Comments and replies |
|
| AI-generated action items |
|
| Emoji reactions |
|
| Video tags |
|
| External references (where the video is shared/embedded) |
|
| Confluence meeting notes URL | |
| Linked Confluence pages | |
| Signed MP4 download URL | |
| Download audio or video to a local file (works when MP4 export is disabled; optional |
|
| All-in-one: metadata + transcript + chapters + summary + comments + tasks |
|
| List your folders | |
| List your workspaces | |
| Space details (name, privacy, primary) | |
| Search folders by name | |
| Folder details | |
| Last timestamp where you stopped watching | |
| Number of videos in your Watch Later list | |
| Total videos created by a user | |
| Your most-used emoji reaction types | |
| Emoji reactions on a specific comment | |
| User profile by ID (name, email, company, avatar) | |
| Search tags in your workspace |
Write (29 tools)
Write tools modify your Loom data. Destructive actions (delete, archive, move) cannot always be undone.
Tool | Description |
| Rename a video |
| Update video description |
| Update video settings (downloads, comments, etc.) |
| Post a comment (with optional timestamp) |
| Edit an existing comment |
| Delete a comment |
| Create an action item on a video |
| Update the content of an action item |
| Delete an action item |
| Mark a task as approved |
| Respond to a task |
| Add an emoji reaction at a timestamp |
| Delete a reaction |
| React to a comment with an emoji |
| Follow/unfollow a video |
| Follow/unfollow a workspace tag |
| Archive or unarchive videos |
| Duplicate a video |
| Permanently delete a video |
| Recover a deleted video from trash |
| Pin or unpin a video in your library |
| Add to Watch Later list |
| Remove from Watch Later list |
| Create a new folder |
| Rename a folder |
| Delete folders |
| Move videos to a different folder |
| Move folders into a different parent folder |
| Share videos to one or more spaces |
Saving to disk
Most per-video read tools accept an optional save_dir parameter. When provided, the tool saves its output to {save_dir}/{video_id}/ and returns the file path alongside the content. When omitted, nothing is saved.
save_dir (and download_media's out_path) must be an absolute path; ~ is expanded. Relative paths are rejected because they would resolve against the server process's working directory, which means nothing for a shared or launchd-managed server.
Just ask naturally — "get the details from my last meeting and save them to ~/Downloads/loom" — and the LLM will pass save_dir="~/Downloads/loom" to the tool.
get_video_details saves each piece individually (metadata.json, transcript.txt, summary.txt, etc.) plus a combined details.md.
Troubleshooting
Auth errors
If you get auth errors, your connect.sid cookie has expired (~30 days). Grab a fresh one from your browser using the steps in Authentication.
Debugging with MCP Inspector
npx @modelcontextprotocol/inspector uvx --from git+https://github.com/karbassi/loom-mcp.git loom-mcpContributing
Issues and pull requests are welcome on GitHub.
License
Available Tools
60 toolsadd_comment_reactionBIdempotent
Add an emoji reaction to a comment on a Loom video.
Inputs: comment_id (required), reaction (required), comment_type. Returns: result
add_reactionA
Add an emoji reaction to a Loom video at a specific timestamp.
Use get_frequent_reactions to discover valid reaction type values.
Inputs: video_id (required), time (required), reaction_type (required). Returns: result
add_to_watch_laterAIdempotent
Add a Loom video to your Watch Later list.
Inputs: video_id (required), minutes_from_utc. Returns: result
approve_taskAIdempotent
Mark an action item (task) as approved on a Loom video.
Inputs: task_id (required). Returns: result
archive_videosBDestructiveIdempotent
Archive or unarchive Loom videos. Archived videos are hidden but not deleted.
Inputs: video_ids (required), archive. Returns: result
create_commentA
Post a comment on a Loom video. Each call creates a new comment.
Inputs: video_id (required), content (required), timestamp. Returns: result
create_folderA
Create a new Loom folder. Each call creates a new folder.
Inputs: name (required). Returns: result
create_taskA
Create an action item (task) on a Loom video. Each call creates a new task.
Inputs: video_id (required), content (required), timestamp. Returns: result
delete_commentADestructiveIdempotent
Delete a comment from a Loom video. This cannot be undone.
Inputs: comment_id (required). Returns: result
delete_foldersBDestructiveIdempotent
Delete one or more Loom folders. This cannot be undone.
Inputs: folder_ids (required). Returns: result
delete_reactionADestructiveIdempotent
Delete an emoji reaction from a Loom video.
Inputs: reaction_id (required). Returns: result
delete_taskADestructiveIdempotent
Delete an action item (task) from a Loom video. This cannot be undone.
Inputs: task_id (required). Returns: result
delete_videoADestructiveIdempotent
Permanently delete a Loom video. This cannot be undone.
To temporarily remove a video, use archive_videos instead. To recover a recently deleted video, use recover_video.
Inputs: video_id (required). Returns: result
duplicate_videoA
Duplicate a Loom video. Creates a new copy each time.
Inputs: video_id (required). Returns: result
edit_commentADestructiveIdempotent
Edit an existing comment on a Loom video. Overwrites the comment text.
Inputs: comment_id (required), video_id (required), content (required). Returns: result
get_backlinksARead-onlyIdempotent
Get external references (backlinks) to a Loom video — where it's been shared or embedded.
Inputs: video_id (required), save_dir. Returns: result
get_captionsARead-onlyIdempotent
Get WebVTT captions of a Loom video with start and end timestamps per cue.
Ideal for precise timing analysis. For plain text with speaker names, use get_transcript instead.
Inputs: video_id (required), save_dir. Returns: result
get_chaptersARead-onlyIdempotent
Get AI-generated chapter markers with timestamps for a Loom video.
Useful for navigating long videos. For a narrative summary, use get_summary.
Inputs: video_id (required), save_dir. Returns: result
get_comment_reactionsARead-onlyIdempotent
Get emoji reactions on a specific comment.
Inputs: comment_id (required), comment_type. Returns: result
get_commentsARead-onlyIdempotent
Get comments on a Loom video, including threaded replies and timestamps.
Inputs: video_id (required), save_dir. Returns: result
get_confluence_pagesBRead-onlyIdempotent
Get Confluence pages linked to a Loom video.
Inputs: video_id (required). Returns: result
get_descriptionARead-onlyIdempotent
Get the AI-generated description of a Loom video with timestamped sections and bullet points.
More detailed than get_summary. For a brief 1-2 sentence summary, use get_summary instead.
Inputs: video_id (required), save_dir. Returns: result
get_download_urlARead-onlyIdempotent
Get a signed download URL for the MP4 file of a Loom video. The URL is temporary and will expire. If no URL is available, call regenerate_mp4 then retry this tool after ~30 seconds.
Inputs: video_id (required). Returns: result
get_folderARead-onlyIdempotent
Get details of a Loom folder including name, visibility, and creator.
Inputs: folder_id (required). Returns: result
get_frequent_reactionsARead-onlyIdempotent
Get your most frequently used emoji reaction types.
Also useful to discover valid reaction type values for add_reaction.
Inputs: no named parameters. Returns: result
get_key_takeawaysARead-onlyIdempotent
Get AI-generated key takeaways from a Loom video as a bullet list.
For a narrative summary, use get_summary. For a detailed timestamped breakdown, use get_description.
Inputs: video_id (required), save_dir. Returns: result
get_last_watch_timeARead-onlyIdempotent
Get the last timestamp (in seconds) where you stopped watching a Loom video.
Inputs: video_id (required). Returns: result
get_meeting_notesARead-onlyIdempotent
Get the Confluence meeting notes URL linked to a Loom video.
Inputs: video_id (required). Returns: result
get_reactionsARead-onlyIdempotent
Get emoji reactions on a Loom video, including who reacted and at what timestamp.
Inputs: video_id (required), save_dir. Returns: result
get_spaceARead-onlyIdempotent
Get details of a Loom space including name, privacy level, and whether it's the primary space.
Inputs: space_id (required). Returns: result
get_summaryARead-onlyIdempotent
Get the AI-generated summary of a Loom video (1-2 concise sentences).
For a detailed timestamped breakdown with bullet points, use get_description. For key highlights as a bullet list, use get_key_takeaways. For chapter markers, use get_chapters.
Inputs: video_id (required), save_dir. Returns: result
get_tagsBRead-onlyIdempotent
Get tags on a Loom video.
Inputs: video_id (required), save_dir. Returns: result
get_tasksARead-onlyIdempotent
Get AI-generated action items (tasks) from a Loom video, including assignee, status, and timestamp.
Inputs: video_id (required), save_dir. Returns: result
get_total_videos_countARead-onlyIdempotent
Get the total number of videos created by a user.
Inputs: user_id (required). Returns: result
get_transcriptARead-onlyIdempotent
Get the full transcript of a Loom video with timestamps and speaker names.
Returns plain text with one line per phrase. For precise start/end timing per cue in WebVTT format, use get_captions instead.
Inputs: video_id (required), save_dir. Returns: result
get_userARead-onlyIdempotent
Get a Loom user's profile by their ID — name, email, company, and avatar.
Inputs: user_id (required). Returns: result
get_videoARead-onlyIdempotent
Get metadata for a Loom video including name, duration, owner, views, and creation date.
For comprehensive info including transcript, chapters, summary, and comments in one call, use get_video_details instead.
Inputs: video_id (required), save_dir. Returns: result
get_video_detailsARead-onlyIdempotent
Get all available information for a Loom video in one call: metadata, transcript, chapters, summary, tasks, and comments.
Use this when you need a complete picture of a video. For just metadata, use get_video.
Inputs: video_id (required), save_dir. Returns: result
get_watch_later_countARead-onlyIdempotent
Get the number of videos in your Watch Later list.
Inputs: no named parameters. Returns: result
list_foldersARead-onlyIdempotent
List your Loom folders, sorted by most recent.
Inputs: limit. Returns: result
list_spacesARead-onlyIdempotent
List your Loom spaces (workspaces).
Inputs: no named parameters. Returns: result
list_videosARead-onlyIdempotent
List your Loom videos, sorted by most recent.
Returns video IDs and names. Use get_video for metadata on a specific video, or get_video_details for comprehensive info including transcript and comments.
Inputs: limit. Returns: result
move_foldersBDestructiveIdempotent
Move one or more Loom folders into a different parent folder.
Inputs: folder_ids (required), destination_folder_id (required). Returns: result
move_videosADestructiveIdempotent
Move one or more Loom videos to a different folder.
Inputs: video_ids (required), folder_id (required). Returns: result
pin_videoAIdempotent
Pin or unpin a Loom video in your library.
Inputs: video_id (required), pinned. Returns: result
recover_videoAIdempotent
Recover a deleted Loom video from the trash.
Inputs: video_id (required). Returns: result
regenerate_mp4A
Trigger MP4 regeneration for a Loom video. Use this when get_download_url reports that no URL is available. Regeneration is asynchronous — after triggering, wait ~30 seconds then retry get_download_url.
Inputs: video_id (required). Returns: result
remove_from_watch_laterADestructiveIdempotent
Remove a Loom video from your Watch Later list.
Inputs: video_id (required). Returns: result
rename_folderBDestructiveIdempotent
Rename a Loom folder.
Inputs: folder_id (required), name (required). Returns: result
respond_to_taskCIdempotent
Respond to an action item (task) on a Loom video.
Inputs: task_id (required), responded. Returns: result
search_foldersARead-onlyIdempotent
Search your Loom folders by name.
Inputs: query (required). Returns: result
search_videosARead-onlyIdempotent
Search Loom videos by keyword. Returns matching video IDs and names.
Inputs: query (required), limit. Returns: result
search_workspace_tagsBRead-onlyIdempotent
Search for tags in your Loom workspace.
Inputs: query (required). Returns: result
toggle_followingAIdempotent
Follow or unfollow a Loom video to get notifications.
Inputs: video_id (required), follow (required). Returns: result
toggle_following_tagAIdempotent
Follow or unfollow a tag in your Loom workspace to get notifications.
Inputs: tag (required), follow (required). Returns: result
update_taskADestructiveIdempotent
Update the content of an action item (task) on a Loom video.
Inputs: task_id (required), content (required). Returns: result
update_video_descriptionADestructiveIdempotent
Update the description of a Loom video. Overwrites the existing description.
Inputs: video_id (required), description (required). Returns: result
update_video_nameADestructiveIdempotent
Rename a Loom video. Overwrites the existing name.
Inputs: video_id (required), name (required). Returns: result
update_video_settingsADestructiveIdempotent
Update settings on a Loom video such as download_enabled, comments_enabled, etc.
Inputs: video_id (required), settings (required). Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 60 tools
Most tools target distinct aspects of Loom videos (reactions, comments, tasks, folders, etc.), and descriptions clarify differences. However, there is potential confusion between similar tools like add_comment_reaction vs add_reaction, and multiple get_* tools for video details (e.g., get_description, get_summary, get_key_takeaways) require careful reading to differentiate.
The naming follows a consistent verb_noun pattern (e.g., create_comment, delete_video, get_transcript) with underscores throughout. Verbs like get, list, search, create, delete, update, archive, rename are used predictably, making the tool set easy to navigate.
With 60 tools, the count is excessively high for a typical MCP server. While the domain is rich, many tools could be consolidated (e.g., multiple get_* variants for video details). The sheer number overwhelms agents and makes selection harder than necessary.
The tool set covers nearly all aspects of video management: CRUD for videos, comments, tasks, folders, spaces, reactions, watch later, pinning, archiving, recovering, duplicating, sharing, and more. It also includes advanced features like backlinks, Confluence integration, and caption retrieval. No obvious gaps for the intended purpose.
Maintenance
Related MCP Connectors
MCP server for Clipkit — gives AI agents a video toolbox via the Clipkit schema.
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceA MCP server that exposes GraphQL schema information to LLMs like Claude. This server allows an LLM to explore and understand large GraphQL schemas through a set of specialized tools, without needing to load the whole schema into the context46MIT
- AlicenseAqualityAmaintenanceMCP server for video analysis — extracts transcripts, key frames with OCR, and annotated timelines from video URLs. Supports Loom and direct video files (.mp4, .webm). Zero auth required.1181,171 npm130MIT
- AlicenseAqualityCmaintenanceA code-mode MCP server for the Unraid 7.2+ GraphQL API that exposes search and execute tools, allowing LLM agents to introspect and call any GraphQL field via sandboxed JavaScript.21MIT
- FlicenseNot gradedqualityNot gradedmaintenanceExtracts transcripts, titles, descriptions, and comments from Loom videos via a remote MCP server.-