Model Context Protocol

Create withMotionLabfrom your AI chat

Add MotionLab as an MCP connector in Claude or ChatGPT, then generate video, images, music and 3D assets, upscale to 4K, and check renders without leaving your conversation.

No API keyYour MotionLab creditsRevocable access

Claude
ChatGPT
Claude Code
Cursor
VS Code
Any client

Claude connectorMotionLab MCP

Connect in Claude

Opens Claude's Add custom connector screen with MotionLab and the MCP URL already filled in.

Or set it up by hand

01
Copy the MCP URL
https://www.motionlab.art/api/mcp
02
Open Claude settings
Go to Settings → Connectors, choose Add custom connector, and paste the URL.
03
Sign in and approve
Connect your MotionLab account, review the permissions, and start creating.

What you can do

From prompt to finished asset

Your assistant can use the MotionLab tools available to your account and keep the entire workflow in one conversation.

Get Motionlab capabilities

List the Motionlab services, models, formats, and credit costs currently available to this account. Call this first to discover what you can create.

Get credit balance

Read the connected Motionlab account's current credit balance. Generations reserve credits up front and refund automatically on failure.

Create input upload URL

Get a short-lived presigned URL for an image, MP4 reference video, or WAV/MP3 reference audio file. Pass the returned owned key to the matching generation input.

Generate image

Create an image from a text prompt (optionally editing uploaded reference images). Spends credits. Returns a job id immediately: retain it, then make separate motionlab_get_job calls until terminal; do not create again while it is pending.

Uses credits

Animate image to video

Animate an uploaded image into a Motionlab video clip. Outputs include an audio track by default; pass an uploaded wav/mp3 as audioKey for Audio-to-Video. Spends credits based on duration and resolution. Returns a job id immediately: retain it, then make separate motionlab_get_job calls until terminal; do not create again while it is pending.

Uses credits

Generate video

Generate video with any currently enabled MotionLab or provider video model. MotionLab accepts prompt-only generation when motionlab_get_capabilities reports requiresStartImage=false and controls.textToVideo=true. The model-specific modes and inputs are returned by motionlab_get_capabilities.

Uses credits

Generate audio

Create a Lyria music track with optional structured music controls and an inspiration image.

Uses credits

Create 3D asset from image

Generate a textured 3D asset (GLB) from an uploaded image. Spends credits. Returns a job id immediately: retain it, then make separate motionlab_get_job calls until terminal; do not create again while it is pending.

Uses credits

Get job status

Check one generation job. If status is PENDING or PROCESSING, call this tool again after 5-10 seconds; continue until a terminal status. On COMPLETED, present outputUrl to the user and render image output inline when supported.

List jobs

List your recent Motionlab generation jobs, newest first. Useful for reusing a previous asset as an input or checking several jobs at once.

66 models, one connection

Everything MotionLab can render

Your assistant reads this catalog live, with the prices below, and picks the right model for what you asked for. No API keys, and nothing to install per model.

MotionLab video engines

Rendered on MotionLab's own infrastructure, with the full control set.

Motionlab FastFast
720P · 1080P · up to 20s · text or image start · seamless loop, end frame, motion reference, audio to video
2–28 credits

Partner video models

Third-party engines, billed from the same MotionLab balance.

Image models

Generate from a prompt, or edit and restyle images you upload.

GPT Image 2.5 SunburstCreate + edit
up to 16 reference images
1 credit
GPT Image 2.5 FlareCreate + edit
up to 16 reference images
1 credit
Grok Imagine Image 2.0Create + edit
up to 3 reference images
1 credit
GPT Image 2Create + edit
up to 4 reference images
1 credit
Nano Banana 2Create + edit
up to 1 reference image
1 credit
FLUX 2 ProCreate + edit
up to 8 reference images
1 credit
Seedream V5 ProCreate + edit
up to 10 reference images
1 credit
Seedream V5 LiteCreate + edit
up to 10 reference images
1 credit

Music models

Structured control over genre, mood, key, BPM, and lyrics.

Lyria 3
30s
1 credit
Lyria 3 Pro
30–180s
2 credits

3D models

Image to 3D. Returns a textured GLB you can open anywhere.

Hi3D Multiview
GLB output
10 credits
Hi3D V3.0
GLB output
39 credits

Try these first

Ask for it in plain language

Once MotionLab is connected, you do not name tools or model ids. Describe the shot and the assistant picks the model, starts the render, and comes back with the file.

Concept to finished shot

“Use MotionLab to generate a matte black espresso machine on wet concrete in 16:9, then animate the best one into a 5 second 1080p clip with a slow push-in.”

Animate a photo you already have

“Here's a product photo. Upload it to MotionLab and animate it into a 6 second clip with a slow orbit and ambient sound, then send me the MP4.”

Score it

“Write me a 30 second cinematic synth track in A minor at 90 BPM for that clip, and tell me what it cost in credits.”

Make it 3D

“Turn this sneaker photo into a textured 3D asset with MotionLab and give me the GLB download link.”

Finish at 4K

“Upscale my last MotionLab render to 4K, and tell me the credit cost before you start.”

Check before you spend

“What can MotionLab make for me right now, which video models are available, and how many credits do I have left?”

Built to stay in control

Your account. Your permissions.

One balance, everywhere

Your assistant uses your MotionLab account and credit balance, at website prices. Failed generations are refunded automatically. See pricing.

Ready for long renders

Video, 3D and upscale jobs continue in MotionLab's render queue while the assistant checks progress and returns the finished asset.

Secure OAuth access

Review the requested permissions before connecting. No MotionLab infrastructure secrets are ever shared, and access can be revoked at any time.

Questions

Everything you need to connect

Still deciding? Browse what MotionLab makes on the home page, the Upscale Lab, or the pricing page.

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external tools. MotionLab's remote MCP server gives an assistant access to image, video, music, 3D and upscaling after you approve the connection.

Which AI apps can connect to MotionLab?

Any client that supports remote MCP servers over Streamable HTTP with OAuth sign-in. This page has step-by-step setup for Claude, ChatGPT, Claude Code, Cursor and VS Code, and the same URL works in other MCP clients.

How do I add MotionLab to ChatGPT?

Turn on developer mode under Settings → Connectors → Advanced, choose Add custom connector, and paste the MotionLab MCP URL from this page. Custom connectors require a paid ChatGPT plan on the web app, and on Business or Enterprise workspaces an admin enables developer mode first.

Do I need an API key?

No. Sign in with your normal MotionLab account and approve the requested access in the browser. There are no keys to copy, and you can revoke the connection at any time.

Which models can the assistant use?

The full catalog on this page: MotionLab Fast and Max, partner video models, image models with editing, music generation, image-to-3D, and upscale. The assistant reads the list live, so it only ever offers what your account can actually run.

How do credits work?

Generations use the same MotionLab account and credit balance as the website, at the same prices. Credits are reserved when a job starts and refunded automatically if the job fails. Your Claude or ChatGPT subscription is not charged for MotionLab renders.

How long do generations take?

Images usually finish in seconds. Video, 3D and upscale jobs are queued and rendered asynchronously, so the assistant can start a job, check its status, and return the finished asset when it is ready.

Can the assistant use my own images and video?

Yes. It requests a short-lived upload URL, sends your file straight to MotionLab storage, and passes the resulting key to a generation tool. Uploads are scoped to your account and are never shared with other users.

How do I disconnect it?

Remove the connector in your AI app, or revoke it from your MotionLab account. Revocation takes effect immediately, and existing tokens stop working on the next call.

Create withMotionLabfrom your AI chat

Copy the MCP URL

Open Claude settings

Sign in and approve

From prompt to finished asset

Get Motionlab capabilities

Get credit balance

Create input upload URL

Generate image

Animate image to video

Generate video

Generate audio

Create 3D asset from image

Get job status

List jobs

Everything MotionLab can render

MotionLab video engines

Partner video models

Image models

Music models

3D models

Ask for it in plain language

Concept to finished shot

Animate a photo you already have

Score it

Make it 3D

Finish at 4K

Check before you spend

Your account. Your permissions.

One balance, everywhere

Ready for long renders

Secure OAuth access

Everything you need to connect