Developers·Model Context Protocol

Connect your AI agent
to Montage

Give Claude, ChatGPT, Cursor, or any MCP-compatible agent the power to ingest video, cut publish-ready clips, and query transcripts — right inside your workflow. Two commands to set up, no API key required.

quick start
npx skills add ukumi-ai/montage-skills
claude mcp add --transport http montage https://mcp.montage.app/mcp
01Install

Install Montage skills

Montage publishes agent skills on skills.sh. Install everything with one command on Claude Code and any skills.sh-compatible agent:

shell
npx skills add ukumi-ai/montage-skills
ingest-video

Ingest a video into Montage from a Google Drive or public URL and track the ingestion pipeline through every step.

create-moment

Create an editable moment (clip) from an ingested video and get back an editor link ready to share.

If your agent doesn't support skills.sh, browse the skill files on the skills page and add them manually.

02Connect

Connect the MCP server

Point your MCP client at the remote server (Streamable HTTP). Authentication is OAuth 2.1 with PKCE and dynamic client registration — no API key needed. Your client opens a browser sign-in automatically.

server url
https://mcp.montage.app/mcp
corpus:read — requiredmoments:readmoments:write
Claude Code (CLI)

Add the server with a single command:

shell
claude mcp add --transport http montage https://mcp.montage.app/mcp
Claude.ai / Claude Desktop

Settings → Connectors → Add custom connector, then paste the server URL:

url
https://mcp.montage.app/mcp
ChatGPT & other MCP clients

Add a remote MCP server with the URL below and complete the OAuth sign-in when prompted:

url
https://mcp.montage.app/mcp
03Reference

Available MCP tools

Once connected, your agent can call these tools directly. Skills orchestrate them for you, but they're all available for custom workflows.

Projects & content
list_projects

List projects owned by the authenticated user.

get_project

Get details for a single project by ID.

get_transcript

Get transcript segments for a project.

get_chapters

Get AI-generated chapters for a project or file.

get_binary_labels

Get utterance binary labels for a project.

Ingestion
upload_video

Import a video from a Google Drive or public URL into a new project.

get_upload_status

Get the progress of an upload_video pipeline.

get_workflow_status

Get the status of a workflow by workflow ID.

Moments & corpus
create_moment

Create a moment (clip) from [start, end] second pairs of an ingested video.

get_moments

Get moments for a project or file, including a gallery view.

get_video_corpus

Get the long-footage video corpus index for a project or file.

query_video_corpus_range

Drill into one time window of a video corpus.

Playbooks
list_playbooks

List playbooks owned by the authenticated user.

create_playbook

Create a new playbook from a natural-language description.

update_playbook

Update an existing playbook via a change description.

get_playbook

Get the current approved playbook by name.

Utility
ping

Confirm the server is reachable and tool calls work.

04Verify

Verify the setup

A quick three-step check to confirm everything is wired up correctly.

1
Call ping

It should return a success response, confirming the server is reachable and your session is authenticated.

2
Call list_projects

It should return the user's Montage projects. An empty list is expected for brand-new accounts.

3
Try a skill end-to-end

Ingest a video with ingest-video, then turn a highlight into a clip with create-moment.

Resources
skills.sh
Montage skills
For AI & LLMs
Product overview for LLMs
Email
Support