Avatrix API · MCP and REST v1
Connect Avatrix to Claude, Cursor or your own product and generate AI influencers, realistic images and UGC videos — without leaving your workflow. In the agent, you don't even need a token.
Uses the credits from your Avatrix plan.
Paste the MCP URL into the agent. Connect opens the Avatrix login (Google or email) and a screen that tells you what the agent will be able to do. The agent gets all of Avatrix: catalog and quotes, upload, library, characters, image and editing, video, Marketing Studio, voice, batch and the AI Chat with your agents.
On claude.ai, in 3 steps
https://avatrix.io/mcpClaude Code: claude mcp add --transport http avatrix https://avatrix.io/mcp
Cursor: Settings → MCP → Add → paste the same URL.
ChatGPT: Settings → Connectors → add the same URL (account with custom connectors enabled).
No token needed
No key to copy, paste or store. You sign in with the Avatrix account you already have and use the credits you've already paid for.
The cost is quoted up front
Every generation tool accepts get_cost: true and returns the price without generating or charging anything. Ask for the quote, then tell it to generate.
n8n and your own backend still use the Bearer av_ from the REST API, right below.
Every tool becomes a call. No button clicking.
"Create 8 photos of Valentina: brunette, 24, light eyes, varied expressions, clean background. Same face in all of them."
Generates the persona with a locked face — the 8 photos look like they were taken the same day, of the same person.
POST /generate/influencer"Vertical photo of Valentina holding the product, clean kitchen background, natural light, lifestyle style."
Places the model in the scene with the product — no photo shoot, no production company.
POST /generate/image"15s vertical video: Valentina talks about the product, natural tone like a review."
Generates the UGC with the persona you defined, ready to run as an ad.
POST /generate/video-ugc"Generate 30 variations of Valentina's profile photo: different angles, emotions and backgrounds. Identical face."
Produces the full batch — you pick the best ones.
POST /generate/influencer (batch)"Animate this photo of Valentina into a 5s video: smooth motion, looking at the camera, no cuts."
Turns the still image into a fluid video — no camera, no set.
POST /generate/video-creativeThe API has no separate billing; each capability follows the plan and the published rollout.
AI influencer
Consistent persona: locked face across multiple photos
Image generation
Realistic photo, product, editorial, lifestyle
UGC video
Vertical video with an AI persona, natural tone
Creative video
Photo animation, persona video in motion
Image editing
Adjust an existing image with a text instruction
Background removal
Automatic transparent background
Skin enhancement
Realistic refinement without overediting
Upscale
Increase resolution without losing detail
There's no separate account on another platform. You connect with the Avatrix account you already have, use the credits from the plan you've already paid for — and the credits never expire.
Base URL: https://avatrix.io/api/v1
Auth: Bearer token (av_...) — generate it at /account/tokens
curl -X POST https://avatrix.io/api/v1/generate/image \
-H "Authorization: Bearer av_seu_token" \
-H "Content-Type: application/json" \
-d '{"prompt": "foto realista de influencer, fundo clean, luz natural"}'Response 202 with a taskId — generation is asynchronous. Poll GET /tasks/{taskId} or set up a webhook.
Do I need a token to connect in Claude?
No. With MCP you paste the URL, log in to Avatrix and click Allow — the agent starts using your account and your credits, with no key to copy or store. An av_ token is only needed for the REST API (n8n, your own backend), and you generate it at /account/tokens.
Which generations are available through the API?
AI influencer, image, UGC video, creative video, image editing, background removal, skin enhancement and upscale. The API catalog is the same as the app's.
How do I know the cost before generating?
You ask before you spend. Every MCP generation tool accepts get_cost: true and returns only the quote, without creating or charging anything — and get_credits shows the account balance. The price is the same as in the app, with no API surcharge. For REST, the cost per endpoint is at /api/docs.
How does async generation work?
You send a POST to the generation endpoint and get back a taskId. Then you call GET /tasks/{taskId} to check the status — or set up a webhook to receive the result when it's ready.
Can I integrate Avatrix into my own product and offer it to clients?
Yes. The API is built for that: you connect, generate through your own workflow and deliver to your client. The terms are at /termos.
Keep reading