Arcframe
MCP — connect your AI

Turn your AI into a creative engine

Connect Arcframe to Claude, ChatGPT, Cursor, Perplexity and any MCP-capable agent — and generate video, images, audio and 3D right inside your conversations. One-click sign-in, no API keys to copy.

Your agent
1

Open connector settings

In the Claude desktop or web app, open Settings → Connectors. Choose “Add custom connector”.
2

Paste the Arcframe URL

Name it Arcframe and paste this URL:
https://mcp.arcframe.ai/mcp
3

Connect & sign in

Click Connect, sign in with your Arcframe account and approve access. Then just ask it to generate something.

Generations use your Arcframe credits and plan limits — exactly like the web app.

Just describe it, Arcframe builds it

A look at the tools your agent can call — and what each one returns when you ask.

Generate video

Make a 6-second cinematic product reveal for our new wireless earbuds, golden-hour lighting.

Generate image

Create a 4K hero image of a matte-black smartwatch on a marble surface, soft studio light.

Generate audio

Write 20 seconds of upbeat lo-fi background music for a startup promo.

Generate 3D

Turn this sneaker photo into a downloadable 3D model I can drop into Blender.

Make a 6-second cinematic product reveal for our earbuds.
Arcframe

Arcframe

On it — calling generate_video with Seedance 2.0. Here’s your job:

videoGenerating · 6s · 1080p

Returned a job ID instantly — I’ll poll get_job and share the finished clip. It’ll cost 12 credits.

videogenerate_video

Script, ad, UGC, news & motion-graphics video from a prompt

imagegenerate_image

Text-to-image, edit, upscale, remove background

audiogenerate_audio

Music, sound effects, speech & voice cloning

3dgenerate_3d

Text or images → downloadable 3D models (GLB)

readget_job

Poll a generation and fetch its result URL

readlist_my_creations

Browse your recent generations

readget_balance

Check your credit balance and plan

readlist_models

See every model with credit costs

The latest models, all in one place

One connection, every frontier model — pick by name, set parameters, generate.

B

Seedance 2.0

ByteDance

G

Google Veo 3.1

Google

K

Kling 2.5 Turbo

Kuaishou

O

Sora 2 Pro

OpenAI

B

Flux 1.1 Pro Ultra

Black Forest Labs

G

Nano Banana 2

Google

O

GPT Image 1.5

OpenAI

S

Stable Audio 2.5

Stability AI

E

ElevenLabs v3

ElevenLabs

M

Meshy v6

Meshy

…and 10+ more across video, image, audio and 3D. Ask your agent to list_models.

A complete creative studio inside your agent

Image and video generation, audio production and 3D — all in a single connection.

Generative media

Image & video generation

From a single still to full cinematic video — choose your model, set the parameters and generate. Flux, Nano Banana, Imagen and GPT Image for stills; Seedance, Kling, Veo and Sora for video. Up to 4K, any aspect ratio, any supported duration.

Generate a cinematic portrait of a woman alone in an empty subway.
Arcframe

Arcframe

Here she is — dark coat, soft fill, standing still in an empty station.

imagePortrait · 4K
videoAnimate →
Score this 20s clip and turn my product photos into a 3D model.
Arcframe

Arcframe

Generating an original track and a textured mesh:

audioLo-fi · 20s
3dGLB · textured
Audio & 3D

Sound and dimension

Generate original music, cinematic sound effects, lifelike speech in 70+ languages and voice clones — then turn text or photos into downloadable GLB 3D models. The same agent, the same conversation, every modality.

One prompt, every modality

Chain tools in a single conversation — your agent generates across video, image, audio and 3D, then hands you the files.

Build a launch kit for our earbuds: a hero image, a 6s ad, and a music bed.
Arcframe

Arcframe

Three generations queued. I’ll deliver download links as each completes:

imageHero · 4K
videoAd · 6s
audioMusic · 20s

All saved to your Arcframe library — browse anytime with list_my_creations.

Got any questions left?

The things people ask most before connecting.

Through the Model Context Protocol (MCP). You add the Arcframe connector URL once inside your agent, sign in with your Arcframe account via secure OAuth, and the agent gains a set of generation tools it can call during your conversation.

Any MCP-capable agent. We've confirmed Claude (desktop & web), and the same standard connector URL works for ChatGPT, Cursor, Perplexity and other MCP clients.

20+ models across all four modalities — Seedance, Veo, Kling, Sora and Wan for video; Flux, Imagen, Nano Banana, GPT Image and Ideogram for image; Stable Audio, Mirelo, ElevenLabs and MiniMax for audio; and Meshy, Hunyuan and Trellis for 3D. Ask list_models to see live credit costs.

Video, images, audio (music, sound effects, speech and voice clones) and downloadable 3D models — all generated from your conversation and saved to your Arcframe account.

No. One-click OAuth signs you in — there are no keys to copy, paste or rotate. You can revoke any connected app anytime in Settings → Apps.

Generations spend your Arcframe credits and respect your plan limits, exactly like the web app. Connecting an agent is free; you only spend credits when you generate.

Generation is asynchronous: a generate_* tool returns a job ID instantly, then your agent polls get_job for the result URL. Images land in seconds; video takes longer depending on the model and duration.

Yes. Each connection is OAuth 2.1 and acts only as you — it can only ever see your generations and spend your credits, never another user's. Tokens are revocable instantly from your settings.

Explore more of Arcframe

Everything your agent can make is also a click away in the app.

Your AI stack just leveled up

Connect Claude, ChatGPT, Cursor or Perplexity to Arcframe in seconds. Video, images, audio and 3D — on demand, from your conversation.

Create a free account

Already connected? Manage connected apps →