muse-image-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MUSE_MODEL | No | Global default model ID. | muse-image-1.0 |
| MUSE_API_KEY | Yes | API key. Without it the server exits immediately and explains itself on stderr. | |
| MUSE_BASE_URL | No | API base URL. | https://api.meta.ai/v1 |
| MUSE_OUTPUT_DIR | No | Output directory, created if missing. | <cwd>/generated-images |
| MUSE_TIMEOUT_MS | No | Per-request timeout in milliseconds. | 120000 |
| MUSE_EXTRA_PARAMS | No | JSON object string — global default extra parameters. | {} |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| generate_imageA | 以 Meta Muse 模型從文字描述生成圖片。圖片會存到本機並回傳絕對路徑（不回傳圖片內容本身，需要看圖請用檔案讀取工具開啟該路徑）。size 參數是長寬比而非精確像素解析度。每張圖片成本 US$0.01。 |
| edit_imageA | 以 Meta Muse 模型依既有圖片與指令生成新圖。images 可填本機檔案路徑或 http(s) 網址，本機檔案會自動轉成 base64，呼叫端不需自行處理。結果存到本機並回傳絕對路徑。每張圖片成本 US$0.01。 |
| iterate_imageA | 以對話方式多輪迭代修改圖片。回傳中一定包含 response_id；下一輪修改時把它填入 previous_response_id 即可延續同一段對話，本 server 不保存任何對話狀態。每張圖片成本 US$0.01。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
generate_image (text-to-image) is clearly distinct, but edit_image (single-turn edit from existing images) and iterate_image (multi-turn conversational edit) overlap in purpose. The descriptions do clarify the difference via the response_id/previous_response_id mechanic, so an agent can likely distinguish them.
All three tools follow a consistent verb_noun pattern: generate_image, edit_image, iterate_image. The convention is predictable and uniform.
Three tools is on the lean side but appropriate for a focused image generation/editing service. Each tool covers a distinct mode (create, single edit, iterative edit), though the surface feels slightly thin.
The core image workflow (generate, edit, iterate) is covered, and tool outputs return file paths with clear semantics. Missing extras like batch generation, upscaling, or listing past generations are minor gaps an agent can work around.