Livepeer MCP. Manage full video pipelines from your chat client.
Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Livepeer manages your entire decentralized video pipeline. Use this MCP Server to upload assets from external URLs, create live streams, generate short clips from active broadcasts, and monitor real-time viewership metrics—all via conversation with your AI agent.
What your AI agents can do
Create clip
Generates a short video clip from an active live broadcast session using time markers.
Create multistream target
Sets up a single content source to be broadcast across multiple destinations like YouTube or Twitch.
Create room
Establishes a new, private WebRTC multi-participant video room for real-time collaboration.
Creates, updates, or terminates live broadcast sessions and multi-platform distribution targets.
Uploads video files from external URLs (HTTP/IPFS) and allows retrieval or deletion of stored media assets.
Generates specific, timestamped short clips from ongoing live streams or schedules background transcode jobs for raw footage.
Queries current audience numbers and detailed usage statistics (bandwidth, minutes) in real time.
Ask AI about this MCP
Supported MCP Clients
Waiting for input…
Livepeer (Decentralized Video) MCP Server: 34 Tools for Stream Ops
Automate every stage of video delivery. These tools let you manage streams, process assets, and track viewership metrics without ever touching a dashboard.
019e5d2ecreate clip
Generates a short video clip from an active live broadcast session using time markers.
019e5d2ecreate multistream target
Sets up a single content source to be broadcast across multiple destinations like YouTube or Twitch.
019e5d2ecreate room
Establishes a new, private WebRTC multi-participant video room for real-time collaboration.
019e5d2ecreate room user
Adds a specific user identity to an existing WebRTC meeting room.
019e5d2ecreate stream
Initializes and sets up a brand new, dedicated live streaming session.
019e5d2ecreate transcode job
Schedules the processing of a video file into different formats or resolutions for on-demand use.
019e5d2ecreate webhook
Sets up a callback URL to receive real-time event notifications from the Livepeer infrastructure.
019e5d2edelete asset
Removes an existing video asset from your account storage.
019e5d2edelete multistream target
Deactivates and removes a configured multi-platform broadcast target.
019e5d2edelete stream
Permanently shuts down and removes an active live stream session.
019e5d2edelete webhook
Removes a configured webhook subscription endpoint.
019e5d2eget asset
Retrieves all metadata and details for a specific stored video asset.
019e5d2eget multistream target
Fetches the current configuration details of a multi-platform broadcast target.
019e5d2eget playback info
Retrieves necessary URLs and metadata needed to play back any asset or active stream.
019e5d2eget realtime viewership
Queries the current number of viewers connected to a live session in real time.
019e5d2eget session
Retrieves overall details about a specific completed broadcast session.
019e5d2eget session clips
Lists and retrieves all generated video clips associated with a given session ID.
019e5d2eget stream
Fetches the current configuration status and details of an active stream.
019e5d2eget task
Retrieves the status (success/failure) and output of a background job, like transcoding or uploading.
019e5d2eget usage metrics
Calculates total usage metrics, including minutes transcoded and data delivered via CDN.
019e5d2eget viewership metrics
Provides a historical breakdown of audience data by device type, browser, or geographic location.
019e5d2eget webhook
Retrieves the configuration and status of an existing webhook endpoint.
019e5d2elist assets
Lists all video assets currently stored in your account's asset library.
019e5d2elist sessions
Retrieves a list of all past broadcast sessions you have run.
019e5d2elist streams
Lists the status and details of all streams, both active and paused.
019e5d2elist tasks
Retrieves a list of all background processing jobs that have run on your account.
019e5d2erequest asset upload
Requests a direct, temporary URL from Livepeer to begin uploading a new asset securely.
019e5d2estart room egress
Broadcasts the output of an active WebRTC room session out to a standard live stream platform (RTMP).
019e5d2eterminate stream
Forces the immediate shutdown and cleanup of any running live streaming session.
019e5d2eupdate asset
Modifies metadata or settings for an existing video asset without changing its content.
019e5d2eupdate multistream target
Changes the destination details (like keys or URLs) for a configured multi-platform broadcast target.
019e5d2eupdate stream
Modifies settings of an existing live stream, such as its title or bitrate.
019e5d2eupdate webhook
Changes the parameters and destination URL for a configured webhook endpoint.
019e5d2eupload asset via url
Upload an asset using an external URL (HTTP, IPFS, Arweave)
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Livepeer (Decentralized Video), then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Livepeer manages your whole decentralized video pipeline. You're gonna use this MCP Server to handle everything from uploading raw assets and setting up live broadcasts, to generating short clips and tracking viewership numbers—all through conversation with your AI agent.
Asset Ingestion and Management
To start working with video, you gotta get the assets into Livepeer first. You can list every asset already stored in your account's library using list_assets. If you need details on a specific file, get_asset pulls all that metadata for you. For bringing content in, you have two main options: you can use request_asset_upload to get a direct, temporary URL from Livepeer and start the upload process there, or if your asset is already hosted somewhere else (like HTTP, IPFS, or Arweave), you just run upload_asset_via_url.
Once you're done with an asset—whether it's old garbage or something that broke—you can delete it using delete_asset.
Stream Setup and Control
Setting up your stream is where the magic happens. To start a totally new, dedicated live broadcast session, you run create_stream. If you need to shut down an active feed immediately, you use terminate_stream, which forces the cleanup of any running session. For advanced distribution, you can create a multi-platform target using create_multistream_target if you want one source beamed out to places like YouTube or Twitch; you'll then manage those settings by calling update_multistream_target and check their config with get_multistream_target.
To deal with private meetings, you can establish a new WebRTC multi-participant room using create_room, and then add specific users to that meeting via create_room_user. You'll need to update the stream settings—like changing the title or bitrate—by running update_stream, and check the status of any active feed with get_stream or see a list of all your streams (active or paused) using list_streams.
The room you created can even broadcast its output out to a standard live platform (RTMP) by calling start_room_egress. You can also delete an entire multi-platform target setup using delete_multistream_target.
Content Processing & Clipping
Don't let great footage sit there. If you want to pull a short segment out of an ongoing live broadcast, run create_clip, which generates that timestamped clip instantly. For background processing—say, converting your raw footage into multiple formats or resolutions for on-demand use—you schedule the job with create_transcode_job. You can track the progress and see if it succeeded or failed by running get_task against a specific job ID, and you get a list of all past jobs using list_tasks.
When you're done with your content management, remember you can also delete any associated assets via delete_asset.
Real-Time Analytics Retrieval & Monitoring
This is how you see if people are actually watching. To get the live count of viewers connected to a stream right now, use get_realtime_viewership. For historical data on who's watching—like breakdowns by device type or location—you run get_viewership_metrics. You can figure out your overall operational costs and usage numbers (total minutes transcoded, data delivered) using get_usage_metrics.
To check the details of past broadcasts, you list all finished sessions with list_sessions then pull specific info for one session ID using get_session. If that session created clips, you can see a list by running get_session_clips. For persistent notifications about what's happening in your infrastructure (like when an asset finishes processing), you set up a callback URL with create_webhook, monitor its details with get_webhook, and clean it up later using delete_webhook.
Other Tools You Need to Know
To keep everything organized, you can list all past broadcast sessions that have wrapped up by running list_sessions. If you need the necessary URLs or metadata just to play back a specific asset or stream, get_playback_info gives it to you. Finally, if you're managing your webhooks and need to change their destination parameters, use update_webhook, and likewise, if you want to modify an existing multi-platform target's keys or URLs, you run update_multistream_target.
How Livepeer MCP Works
- 1 1. Subscribe to the Livepeer server and input your API Key.
- 2 2. Tell your AI client exactly what needs doing (e.g., 'Start a stream for my new product launch').
- 3 3. The agent executes the necessary tool calls (like
create_stream), managing the entire video infrastructure setup in chat.
The bottom line is you run complex, multi-stage video operations—from ingest to analysis—using simple natural language commands.
Who Is Livepeer MCP For?
This server is for the Video Engineer who spends hours on web dashboards just setting up a simple stream. It's for the Content Creator who needs quick analytics without jumping between tools, and the DApp Developer needing to integrate complex video delivery features into code.
Automates pipeline deployments by calling create_stream or managing multistream targets (create_multistream_target) without manual dashboard clicks.
Quickly generates highlight reels using create_clip directly from a live stream and checks audience performance with get_realtime_viewership.
Integrates decentralized video capabilities into applications by managing assets (upload_asset_via_url) or querying usage data via get_usage_metrics.
What Changes When You Connect
- Instantly get performance data. Instead of manually checking dashboards, call
get_realtime_viewershipto know the audience size and location breakdown instantly. - Automate content processing. Use
create_clipto generate highlight reels from an active stream—no manual recording or timeline scrubbing required. - Handle massive scale without effort. Set up multi-platform delivery using
create_multistream_target, routing your single source feed to multiple services simultaneously. - Track costs and usage accurately. Run
get_usage_metricsanytime to get a precise count of total minutes transcoded and data delivered, simplifying billing audits. - Manage entire rooms via code. Use
create_roomorstart_room_egressto set up private collaboration spaces and then broadcast them live—all within the agent workflow.
Real-World Use Cases
The Post-Event Recap
A content creator just finished a 4-hour webinar. Instead of downloading hours of footage to manually find clips, they tell their agent: 'Generate three 60-second clips from the last hour covering Q&A.' The agent runs create_clip with precise timestamps and delivers the ready-to-use assets.
The Live Incident Response
An engineer detects a sudden drop in viewership. They ask their agent to run get_realtime_viewership. The agent returns data showing that 80% of viewers are on mobile Chrome, leading the engineer to adjust stream settings via update_stream immediately.
Cross-Platform Launch
A marketing team needs a single live event broadcast everywhere. They tell their agent: 'Set up multi-platform delivery for the new product.' The agent runs create_multistream_target, linking one source to both Twitch and YouTube simultaneously.
The Clean Up Crew
A developer finishes a test stream. Instead of manually logging into three different dashboards, they ask their agent: 'Terminate the stream and delete all associated assets.' The agent runs terminate_stream followed by delete_asset, cleaning up everything in one command.
The Tradeoffs
Treating it like a file storage API
Trying to use list_assets and then manually calling get_playback_info for every single asset ID just to see if the video works.
→
If you need to confirm connectivity, get the stream status first by running get_stream. Use get_playback_info only when you have a specific asset or session ID to verify its viewing URLs.
Ignoring background jobs
Running create_transcode_job and then immediately assuming the video is ready, leading to failed playback attempts.
→
Always check the status of your job by calling get_task. Wait until the task reports 'Success' before attempting to use the resulting asset.
Trying to stream a recorded file
Attempting to start live streaming using only an uploaded video URL, which will fail.
→
To broadcast a pre-recorded file as if it were live, you must first run create_transcode_job and then use the resulting output stream ID in your publishing workflow.
When It Fits, When It Doesn't
Use this server if your process involves multiple stages: content ingestion (via upload_asset_via_url), real-time event monitoring (get_realtime_viewership), and continuous management of the broadcast source itself. You need control over both the 'front end' experience—like setting up a room with create_room—and the 'back end' metrics via get_usage_metrics. Don't use this if you just want to send a simple message or list files; for that, a basic messaging tool is sufficient. If your goal is simply retrieving metadata on one asset, stick to get_asset; don't try to manage streaming lifecycle tools like create_multistream_target when all you need is the URL.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Livepeer. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 34 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Setting up a live stream used to mean clicking through five different dashboards.
Today, setting up an event requires navigating to the Stream Manager dashboard. Then, you have to input credentials for Twitch, copy that key into another window, then switch to the CDN console to upload assets, and finally write down a list of all the URLs needed just to play it back. It's manual overhead.
With this MCP server, your agent handles the complexity. You tell it: 'Set up live streaming for my product demo.' The agent executes `create_stream` and handles the initial setup steps, giving you immediate stream keys and necessary endpoints right in the chat.
Livepeer (Decentralized Video) MCP Server lets you manage video ops from chat.
The process of monitoring audience drop-off used to involve running reports, downloading CSVs, and manually cross-referencing device type against location data. It was slow and often out of sync.
Now, your agent runs `get_viewership_metrics` on demand. You get the actionable breakdown—device vs. location—in a single response. You stop reporting metrics; you start reacting to them.
Common Questions About Livepeer MCP
How do I check how many people are watching right now using get_realtime_viewership? +
You call get_realtime_viewership and the agent returns a live count. This query gives you instant metrics, preventing the need to refresh complex dashboard pages.
Can I use create_clip for old streams or only active ones? +
create_clip is designed specifically for generating short videos from an active live stream. For clips from past sessions, you should first retrieve the session details using get_session, and then check if associated clips exist via get_session_clips.
What's the difference between list_assets and get_asset? +
list_assets gives you a directory, showing names and IDs of all stored videos. get_asset requires you to specify one asset ID and returns all its detailed metadata.
How do I make my stream available on multiple platforms? +
You use create_multistream_target. This tool sets up the connection, allowing your content source to automatically broadcast across several different services simultaneously.
When I use `create_stream`, what information does it need besides a title? +
It requires specific parameters like bitrate, codec, and duration settings. You must provide these details so the server can correctly initialize the streaming pipeline.
If I change my delivery platform, how do I update my targets using `update_multistream_target`? +
You need to pass the existing target ID and the new configuration parameters. The tool validates these changes before applying them across all connected endpoints.
How can I check if a background job, like transcoding, finished using `get_task`? +
The function returns a status code that tells you the current state. Look for 'completed' or 'failed'; otherwise, it means the task is still processing.
Before uploading a large file, how do I get an upload URL using `request_asset_upload`? +
The tool immediately generates a secure, temporary URL for you. You use this link to push your data directly, bypassing the need for client-side uploads.
Can I upload videos directly from IPFS or Arweave? +
Yes! Use the upload_asset_via_url tool and provide the decentralized storage URL. Livepeer will ingest the asset and make it available for playback.
How do I create a highlight clip from a running stream? +
Use the create_clip action by providing the streamId and the specific startTime and endTime. The agent will trigger the clipping process on the Livepeer network.
Can I monitor how many viewers are currently watching my stream? +
Absolutely. Use the get_realtime_viewership tool with your playback ID to fetch current audience metrics directly from the Livepeer analytics engine.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Headscale (Tailscale Alternative)
Manage your private Tailscale network via Headscale — control users, nodes, and pre-auth keys directly from your AI agent.
Aliyun VOD / 阿里云视频点播
Alibaba Cloud's professional video pipeline — manage media, retrieve play URLs, and audit transcode templates via AI.
Cisco Meraki
Manage cloud-managed networking via Cisco Meraki — track devices, monitor clients, and audit network health directly from any AI agent.
You might also like
AssessTEAM
Evaluate employee performance, run 360-degree reviews, and track team goals with streamlined assessment workflows.
Deputy
Equip your AI agent to manage employee rosters, track timesheets, and monitor leave requests via the Deputy API.
Laravel Forge
Manage Laravel Forge servers, orchestrate site deployments, and query databases directly from your AI agent.