imagine-mcp

- Multimodal understanding-- Describe, classify, or reason over images and videos (Gemini handles mixed image + video in one call)
- Image generation-- Text-to-image and image-to-image (edit / inpaint) across Gemini Imagen, OpenAI gpt-image, Grok Imagine
- Video generation-- Text-to-video and image-to-video (Gemini Veo 3.1, Grok Imagine Video)
- 3 providers x 2 tiers-- Same interface forgemini/openai/grokatpoor(cheap/fast) orrich(high quality); swap via parameter
- Open model passthrough-- Understanding routes through litellm; pass anyprovider/model, or configure an ordered model chain (no hardcoded catalog)
- Degraded mode-- Server starts with zero credentials and surfaces remaining providers as you add keys
- Response cache-- Disk-based caching ofunderstandresponses with configurable TTL
- Dual transport-- pure stdio with provider env vars (default) or HTTP multi-user with paste-token relay form

# uvx -- recommended, runs the published PyPI package uvx imagine-mcp # Docker docker run -it --rm ghcr.io/n24q02m/imagine-mcp:latest

Install with an AI agent-- paste this to your AI coding agent:

Two transports (defaultstdio; opt intohttpwith--http,MCP_TRANSPORT=http, orTRANSPORT_MODE=http):

- stdio(default) -- single-user, reads credentials from env vars only. Exits if none of the three provider keys are set.
- http-- HTTP daemon. Local self-host on127.0.0.1by default, or multi-user remote (per-JWT-sub credential isolation) whenPUBLIC_URL+MCP_DCR_SERVER_SECRETare set. In HTTP mode credentials are entered through a browser form at/authorize.

All optional -- the server starts in degraded mode and surfaces whichever providers have a key. Set at least one.

When a tool is called without an explicitprovider, the first key present wins in the orderXAI_API_KEY->OPENAI_API_KEY->GEMINI_API_KEY.

Model choice passes straight through to litellm (understand) or the native provider SDK (generate) -- there is no hardcoded model catalog. Each chain is a CSV of litellmprovider/modelentries; the order is the fallback order.

Understanding is routed through litellm (provider/modelpassthrough), so any litellm provider works -- supply that provider's<PROVIDER>_API_KEY. Generation stays on the native provider SDKs (Gemini, OpenAI, Grok). Example:

Trust boundary: when the form is gated by ashared*MCP_RELAY_PASSWORD, the username is a partition key, not a secret -- anyone who knows that password can type any username and reach that bucket. That is fine for a trusted group; an untrusted multi-tenant deployment needs a per-user secret or delegated OAuth instead.

One-time migration:existing users must re-enter their credentials once after this change. Nothing is deleted; credentials stored under the old random subject are simply no longer addressed.

git clone https://github.com/n24q02m/imagine-mcp.git cd imagine-mcp mise run setup # or: uv sync --group dev mise run dev # run the server in stdio mode (add --http for the HTTP daemon)

Run your own imagine instance serverless on Cloudflare (Worker + Container + KV). Storage is KV-only -- the per-user credential vault lives in KV, and generation returns base64 only because the container filesystem is ephemeral (IMAGINE_OUTPUT_MODE=base64).

Prerequisites:a Cloudflare account on theWorkers Paid plan-- required for Containers (the Cloudflare free tier does not include Containers) -- and thewranglerCLI.
- git clone https://github.com/n24q02m/imagine-mcp && cd imagine-mcp
- wrangler login
- Create the KV namespace (imagine is KV-only -- no D1 or Vectorize), then paste the returned id intowrangler.jsonc(the<imagine-kv-namespace-id>placeholder):

wrangler kv namespace create imagine-kv

docker pull ghcr.io/n24q02m/imagine-mcp:beta docker tag ghcr.io/n24q02m/imagine-mcp:beta imagine-mcp:beta wrangler containers push imagine-mcp:beta # prints registry.cloudflare.com/<ACCOUNT_ID>/imagine-mcp:beta

wrangler secret put CREDENTIAL_SECRET wrangler secret put MCP_DCR_SERVER_SECRET wrangler secret put MCP_RELAY_PASSWORD wrangler secret put GEMINI_API_KEY # optional provider default wrangler secret put OPENAI_API_KEY # optional provider default wrangler secret put XAI_API_KEY # optional provider default

Thehttpcontainer image already runs multi-user (MCP_TRANSPORT=httpis baked into the image target). Storage maps to Cloudflare viaMCP_STORAGE_BACKEND=cf-kv(encrypted credential vault) withIMAGINE_OUTPUT_MODE=base64, which forces base64 responses so no media path is written to the ephemeral container filesystem.

Hosted remote MCP server for AI image and video generation from one Streamable HTTP endpoint.

Style-locking AI image engine with a hosted MCP server. Define one master style and generate unlimited matching images.

Generate professional fashion & product visuals — images and videos, virtual try-on, product-to-model, image editing and HD upscaling — right inside your chat, using your own account.

AI Video, Image & Audio Generation with over 150 models

Convert photos to LEGO-style brick mosaic previews with AI-powered color quantization. MCP server for Claude, Cursor, and other AI agents.

Official Cannon Studio MCP for AI video, image, 3D, audio, workflow, pricing, model, and developer API guidance.

a specialised AI text remover for photos and graphics

Official FLUX MCP server from Black Forest Labs. Generate, edit, vary, and browse FLUX.2 images directly in any MCP-compatible client

MCP server for OpenAI GPT image generation and editing