Flux AI

npx add-mcp 'https://fluxaipro.art/api/mcp'

MCP serverPreview.

Create images, video and music from the AI tools you already use

Connect Claude, ChatGPT, Cursor or any MCP client to Flux AI. Your assistant picks a model, shows the exact price, waits for your approval, then generates with your credits.

Server URL

https://fluxaipro.art/api/mcp

• Sign in with your Flux AI account. No API key.
• Uses your credit balance. Nothing is charged without approval.
• 18 tools · 20 models

ChatGPTClaudeClaude CodeCursorCodex CLI

How it works

Three steps from a sentence to finished media

1. 01

   Connect once

   Add the server URL to your client and sign in. The client receives its own access, which you can remove at any time.
2. 02

   Describe what you need

   Ask in plain language. The assistant reads each model's real parameters and estimates the cost before anything runs.
3. 03

   Approve and collect

   Confirm the quoted credits. Results arrive as links and, in clients that support it, as a playable gallery.

Connect

Choose your client

Every client uses the same server URL. You sign in in the browser the first time.

1. 1 Open Settings → Apps & Connectors → Advanced settings and turn on Developer mode.
2. Choose Create, name it Flux AI, paste the server URL and pick OAuth authentication.

   https://fluxaipro.art/api/mcp
   

3. 3 Sign in to Flux AI when the browser opens, then enable Flux AI from the + menu in a chat.

Try this first

“List the Flux AI image models, then estimate a 1:1 product photo of a ceramic mug on linen.”

Tools

What your assistant can do

Estimating never charges. Only the generate tools spend credits, and only at the price you approved.

Discover

Browse available models, read a model's exact parameters and check your balance.

• list_media_models
• get_media_model
• get_credit_balance
• list_image_skills
• plan_image_skill

Estimate and generate

Validate a request and quote it, then submit the approved image, video or music job.

• estimate_image_skill
• generate_image_skill
• estimate_image_generation
• generate_image
• estimate_video_generation
• generate_video
• estimate_music_generation
• generate_music

Follow results

Poll status and fetch finished media with final billing.

• get_generation_status
• get_generation_result

Speech

Browse speech voices and play your saved recordings.

• list_speech_voices
• list_speech_recordings

Reference files

Get a private upload link for a reference image, video or audio file.

• create_media_upload

Models

Models available through MCP

This is the same live list your assistant gets from list_media_models.

Image11

• GPT Image 2.5 Flare Fast generation and editing for everyday creativity
• GPT Image 2.5 Sunburst Fine detail and precise editing for complex visuals
• Nano Banana 2.1 1K–4K image generation and reference-photo editing
• GPT Image 2 Edit Precise text, layouts and reference-image editing
• FLUX 3 Image High-resolution images, clear typography and reference-image edits
• Agnes Image 2.5 Flash Fast image generation and reference-image editing
• Flux 2 Klein Fast image generation for everyday creative ideas
• Seedream 5.0 Pro ByteDance image model with generation and editing up to 2K
• Grok Imagine Image 2.0 Edit High-quality xAI image generation and reference editing up to 2K
• Nano Banana Pro Google image model for detailed visuals and reference edits
• Qwen-Image Text-rich layouts, detailed prompts and image editing

Video7

• Seedance 2.5 Up to 30s · 480p / 720p / 1080p · Optional audio
• MiniMax H3 Up to 15s · 768P / 2K · Native audio
• Gemini Omni 1.1 Flash Auto duration · 360p / 720p / 1080p / 4K · Native audio
• Kling 3.0 Motion Control Auto duration · 720p / 1080p · Native audio
• Seedance 2.0 Up to 15s · 480p / 720p / 1080p / 4K · Optional audio
• Seedance 2.0 Fast Up to 15s · 480p / 720p · Optional audio
• Seedance 2.0 Mini Up to 15s · 480p / 720p · Optional audio

Music2

• Suno Create songs from ideas or lyrics, with vocals or instrumental output.
• Flow Music Create music from descriptions or lyrics, with tempo and duration controls.

Billing and access

You stay in control of every credit

Free estimates

Every request is priced before it runs. Estimating never creates a task.

Approval before charging

A generation must carry the approved price. If the price changes, the assistant has to ask again.

No double charges

Retries reuse an idempotency key, so a dropped connection never pays twice.

Disconnect any time

See which apps can use your account and remove access in one click.

Permissions

mcp:read

Browse models, estimate prices, read your balance and results.

mcp:generate

Submit approved generations that use credits.

mcp:upload

Create upload links for reference files.

Questions

What is the Flux AI MCP server?+

A Model Context Protocol endpoint that lets AI assistants use Flux AI's image, video and music models on your behalf, with your account and credits.

Do I need an API key?+

No. Your client signs in with OAuth through your Flux AI account. Each client gets its own access that you can remove from Connected apps.

How are credits charged?+

The same way as on the website. The assistant estimates first, you approve the quoted credits, and failed generations follow the usual refund rules.

Why does opening the server URL in a browser not show a page?+

The URL is for MCP clients, which talk to it with POST requests. Opening it in a browser brings you to this page.

Can I use my own reference images?+

Yes. The assistant creates an upload link, sends the file, and passes the resulting URL to the model.

How do I remove an assistant's access?+

Open Connected apps and choose Disconnect. Access stops immediately; reconnecting requires signing in again.