Luma AI MCP. Control camera motion and video flow directly through your agent.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Luma AI MCP Server lets your agent generate cinematic video and photorealistic images using Luma Dream Machine. You can prompt for text-to-video, animate still photos (`lm.image_to_video`), or control precise camera movements like pan, tilt, dolly, and orbit—all from conversation.
What your AI agents can do
Lm.camera control
Generates videos using Luma Dream Machine while simulating specific camera movements like pan, tilt, dolly, and orbit.
Lm.delete generation
Removes a video generated by Luma Dream Machine.
Lm.extend video
Adds new footage to an existing Luma video, letting the scene continue past its original end point.
Direct AI-generated videos by specifying camera movements like panning, tilting, dollying, or orbiting.
Transform a static image into a dynamic video using Luma Dream Machine's motion coherence engine.
Create cinematic videos directly from natural language prompts using the high-fidelity Ray-2 model.
Smoothly transition between two images (lm.interpolate) or add footage to an existing clip (lm.extend_video).
Rapidly create high-resolution, photorealistic still images using the Luma Photon-1 model.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Luma AI (Generative Video & Creative) MCP Server: 10 Tools
Use these tools to orchestrate complex video creation workflows, from generating initial concepts to adding precise camera movements and extending footage.
019d75calm.camera control
Generates videos using Luma Dream Machine while simulating specific camera movements like pan, tilt, dolly, and orbit.
019d75calm.delete generation
Removes a video generated by Luma Dream Machine.
019d75calm.extend video
Adds new footage to an existing Luma video, letting the scene continue past its original end point.
019d75calm.get credits
Checks how many credits you have left for Luma Dream Machine generation.
019d75calm.get generation
Retrieves the current status (queued, dreaming, completed) and URL for a specific video generation job.
019d75calm.image to video
Turns a single static image into an animated video clip using Luma Dream Machine.
019d75calm.interpolate
Creates smooth, high-quality video transitions between two distinct keyframe images using Luma Dream Machine.
019d75calm.list generations
Lists your recent Luma Dream Machine jobs, showing IDs, prompts, status, and timestamps.
019d75calm.text to image
Generates photorealistic still images based on a text prompt using the Luma Photon-1 model.
019d75calm.text to video
Creates cinematic videos from scratch using a text prompt via Luma Dream Machine, supporting looping options.
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 Luma AI (Generative Video & Creative), 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
Yo, this server lets your agent take full control of cinematic video and photorealistic image generation using Luma's Dream Machine and Photon-1 models. You don’t just get random clips; you tell it exactly how you want the camera to move or what kind of motion you need.
Making Videos Move Like a Pro
- Generating from Scratch: Want a cinematic video? Just give your agent a text prompt, and
lm.text_to_videouses Luma Dream Machine's Ray-2 model to build the whole thing for ya. You can even tell it if you want the final clip to loop—it handles that. - Controlling the Shot: Forget basic video prompts. With
lm.camera_control, your agent doesn't just generate footage; it simulates specific, professional camera movements like panning across a scene, tilting up, dollying in for focus, or orbiting around an object. You nail the shot composition every time. - Bringing Photos to Life: Got a killer still image?
lm.image_to_videouses Luma Dream Machine's motion coherence engine to turn that flat picture into an animated video clip. It adds believable movement without looking fake. - Smooth Transitions: If you have two separate keyframe images,
lm.interpolatemakes them transition together. You get a smooth, high-quality video blend between the two points, making your sequence look professional. - Extending Footage: Did your perfect clip end too soon? Use
lm.extend_video. It takes an existing Luma video and adds brand new footage to the end, letting the scene continue past its original stopping point.
Generating Images and Managing Jobs
- Photorealistic Still Shots: For static images that still need to look insane,
lm.text_to_imageuses the Luma Photon-1 model. Just give it a text prompt, and you get high-resolution, photorealistic stills—no video needed. - Checking Your Work: You can always check what's going on with your generation jobs.
lm.list_generationspulls up a list of all your recent Luma Dream Machine work, showing the job IDs, the prompts you used, when they were created, and their current status. If you need to know where a specific job is,lm.get_generationretrieves its current status (it's queued, dreaming, or completed) along with the final URL. - Cleanup: Made a mistake? Use
lm.delete_generationto wipe out any video generated by Luma Dream Machine that you don't want anymore. - Credits Check: Don’t get caught empty-handed. Run
lm.get_creditsat any time to check how many credits you've got left for the whole shebang.
Basically, your agent calls these tools when you tell it what you need—whether that’s a text prompt or specific camera movements—and the server handles running those complex generation jobs and getting the result back.
How Luma AI MCP Works
- 1 Subscribe to this server and provide your personal Luma AI API Key.
- 2 Use your agent client (Claude, Cursor, etc.) to issue a command (e.g., 'Make a video of X with a dolly shot').
- 3 The agent executes the required tools (
lm.text_to_video+lm.camera_control), and you get back the final MP4 link.
The bottom line is: your AI client acts as the conductor, calling Luma's specific video generation tools so you don't have to write complex code for every step.
Who Is Luma AI MCP For?
This is for visual creators who spend too much time wrestling with manual rendering pipelines. If you're a Video Editor tired of spending hours manually keyframing or color-correcting B-roll, or a Creative Director needing to prototype dozens of storyboards in an afternoon, this server cuts the friction out.
Needs high-quality B-roll and cinematic sequences instantly. They use lm.extend_video to finish a shot or run lm.text_to_video for background material.
Prototypes visual concepts fast, commanding the agent to generate varied styles and camera paths using lm.camera_control without touching After Effects.
Iterates on photorealistic imagery or complex video transitions by chaining lm.text_to_image followed by lm.interpolate for concept art development.
What Changes When You Connect
- You generate high-quality B-roll without manual rendering. Tools like
lm.text_to_videolet you build entire scenes just by describing them in plain language. - Stop using stock footage that looks fake. By running
lm.image_to_video, you take your own concept art and give it believable, consistent motion dynamics. - Need a shot to last 15 seconds but only generated 8? Use
lm.extend_videoto seamlessly tack on more footage without breaking the scene's logic. - Camera movements are where this shines. Instead of flat shots, use
lm.camera_controlto guide your AI camera with precise parameters—pan, tilt, dolly, orbit. - Rapid visual concepting is fast. Use
lm.text_to_imagefor quick mockups orlm.interpolateto smooth the transition between two key concept frames.
Real-World Use Cases
Building a Movie Trailer Sequence
A director needs five shots: 1) A wide shot of a city (using lm.camera_control for an orbit). 2) An animated close-up of a character's face (lm.image_to_video). 3) The transition between the two must be smooth, so the agent calls lm.interpolate. 4) Finally, they need to loop the last shot using parameters in lm.text_to_video.
Creating a Product Background Loop
A marketing team needs an endless background video of smoke or liquid effects. They use lm.text_to_video with the looping option set to true, ensuring the agent manages the parameters and credits using lm.get_credits.
Developing a Concept Art Storyboard
An artist has two rough sketches for scene A and scene B. They use lm.interpolate to bridge those two images into one smooth video segment, then feed that result back into the agent to write a descriptive prompt for the next shot.
Debugging and Managing Jobs
A user submits five complex generation requests. To know which ones are finished or failed, they first run lm.list_generations to get all IDs. Then, they check status one by one using lm.get_generation until everything is 'completed'.
The Tradeoffs
Trying to describe the whole video in one prompt
Prompt: 'Make a 10-second cinematic video of a car driving while panning left, then zooming into a building, and looping it.' This is too complex for a single pass.
→
Break it down. First, use lm.text_to_video with the base prompt. Second, call lm.camera_control to specify the pan/zoom movement in a second message. Third, if you need looping, explicitly tell the agent: 'Now extend that shot and loop it.' Use the tools sequentially.
Forgetting the asset lifecycle
Creating a video, then needing to change one small detail (e.g., changing the time of day). You have to start over.
→
Use lm.text_to_image first to create the perfect still frame for the new moment. Then use lm.image_to_video on that specific image, and finally stitch it back into your timeline using a prompt or an agent workflow.
Ignoring credit management
Running five complex video generations without checking the budget, only to find out halfway through the job that you're over limit.
→
Always start by running lm.get_credits before launching any major generation sequence. This keeps your workflow accountable and prevents unexpected failures.
When It Fits, When It Doesn't
Use this server if you need to move beyond simple, single-shot generative content and require detailed control over the video's physical mechanics—camera movement, scene transitions, or temporal looping. If your project is 90% text-based (e.g., 'Write a script about space travel'), don't use this; just use an LLM. But if you have a concept that needs visual polish, especially cinematic B-roll, then yes, connect it here. Don't try to replicate professional video editing software entirely in the chat—that's impossible. Use lm.camera_control for motion guidance, but if the underlying model can't achieve the look, you're out of luck. This is a powerful toolset for generation, not post-production fixes.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Luma AI. 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 10 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Telling an AI to create video used to mean endless manual prompting and stitching.
Look, before this, if you wanted a smooth transition between two concepts—say, a forest giving way to a desert—you'd generate the first clip, download it, then manually keyframe the second concept, trying to match the lighting and movement. It was tedious, time-sucking work that required specific software knowledge.
Now? You just tell your agent: 'Generate scene A, followed by scene B.' The agent handles calling `lm.interpolate`, figuring out the correct parameters, and delivering a finished video segment. You get the result without touching After Effects.
Luma AI (Generative Video & Creative) MCP Server: Generate complex media streams.
You no longer have to manage your API calls, track job IDs, or guess if the video is ready. The agent takes care of that orchestration—it runs `lm.text_to_video`, then it loops and polls using `lm.get_generation` until the final MP4 link pops out.
This capability means you spend time on the creative idea, not on debugging API calls or waiting for job status updates.
Common Questions About Luma AI MCP
How do I make a video with camera movements using lm.camera_control? +
You specify the movement (e.g., 'dolly shot') in your prompt, and the agent runs lm.camera_control. This directs Luma Dream Machine to calculate the motion parameters for you.
Can I animate a drawing using lm.image_to_video? +
Yes. Simply provide the image URL, and the agent runs lm.image_to_video. It uses that static visual as the initial frame, bringing it to life with AI dynamics.
How do I manage my credits using lm.get_credits? +
Just ask your agent: 'What are my Luma AI credits?' The agent calls lm.get_credits and reports the current balance, so you know when to slow down.
Is there a tool for extending an existing video? Use lm.extend_video. +
Yes, that's exactly what lm.extend_video is for. It takes your current video and adds new footage to the end of it, making sure the scene continues naturally.
How do I generate photorealistic images using the `lm.text_to_image` tool? +
It uses the Luma Photon-1 model to create high-resolution visuals. Simply send a detailed text prompt, and the tool returns static image assets for rapid design iteration.
How do I check if my video generation is finished using `lm.get_generation`? +
You provide your unique Generation ID to this function. It tells you the current status—queued, dreaming, completed, or failed—and provides the final MP4 link when it's ready.
What is the process for creating smooth transitions between scenes with `lm.interpolate`? +
You give the tool two distinct keyframe images. It then generates a seamless, high-quality video that bridges the visual gap and creates a professional transition effect.
If I need to clear out old projects, should I use `lm.delete_generation`? +
Yes, this tool completely removes a specified Luma Dream Machine generation and its associated video file from your account. It's the cleanup function you need.
How do I check if my video generation is finished? +
Use the lm.get_generation tool with the Generation ID provided. Your agent will poll the Luma API and report the current state (queued, dreaming, or completed). Once finished, it will return the final MP4 video URL.
Can I control the camera movement in my AI-generated video? +
Absolutely. Use the lm.camera_control tool. You can provide a scene prompt and a JSON block defining the movement type (e.g., orbit, pan, tilt) and magnitude, allowing for professional cinematographic directing.
Can my agent extend an existing Luma video with more footage? +
Yes. The lm.extend_video tool allows you to provide a continuation prompt and a previous Generation ID. Your agent will trigger Luma to seamlessly expand the scene, maintaining visual and structural consistency.
Multi-server workflows that include Luma AI (Generative Video & Creative) MCP
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Mem0
Give your AI agent persistent memory — store, search, and recall facts, preferences, and context across sessions using the leading agent memory platform.
Cohere (Embed & Rerank)
Empower RAG via Cohere — generate high-quality text embeddings, rerank documents for better accuracy, and perform AI classification directly from any AI agent.
NVIDIA Vision
Generate images, analyze visuals, detect objects, and caption images via NVIDIA Vision APIs.
You might also like
Gusto
Manage employees, run payroll, view benefits and time-off policies — HR automation for AI agents.
Salesforce Service Cloud
Manage support cases, search the knowledge base, track case metrics, and resolve customer issues through natural conversation.
Reflect
Equip your AI to read, write, and explore your networked thought graph in Reflect Notes securely via their API.