What can you do with this server?

You can generate and edit NovelAI images through MCP tools. * Generate text-to-image art with V5/V4.5 models, custom sizes, multi-character prompts, transparent backgrounds, text rendering, comic panels, and furry/background modes. * Redraw existing images with img2img or inpaint masked regions. * Use Director Tools for background removal, lineart, sketch, colorize, emotion changes, and declutter. * Upscale images and perform Vibe Transfer using a reference image’s style and palette. * Suggest Danbooru/NovelAI tags for autocomplete.

How do I use novelai-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@novelai-mcp Generate a portrait of a girl with silver hair and blue eyes in a sunflower field." That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

novelai-mcp

by DesmondFox

Overview Schema Source Related Servers Discussions Admin

Python

Remote

NovelAI MCP Server

Русская версия · Українська версія

An MCP server that lets Claude, Cursor and other MCP clients generate and edit images with NovelAI.

Supports NovelAI Diffusion V5 (Full and Curated) and V4.5: text-to-image, img2img, inpainting, Director Tools, Vibe Transfer, upscaling and tag suggestions.

This is an unofficial project. It is not affiliated with or endorsed by NovelAI / Anlatan.

Tools

Tool	What it does
`novelai_generate_image`	Text-to-image. Size presets or custom size, multi-character prompts with positions, in-image text, transparent background, comic panels, furry and background modes.
`novelai_img2img`	Redraws an existing image from a new prompt. `strength` (0.0–1.0) controls how much changes.
`novelai_inpaint`	Redraws the masked area of an image. White in the mask = repaint, black = keep.
`novelai_augment`	Director Tools: `bg-removal`, `lineart`, `sketch`, `colorize`, `emotion` (24 emotions), `declutter`.
`novelai_upscale`	Neural upscale. NovelAI chooses the factor.
`novelai_vibe_transfer`	Generates a new image in the style and palette of a reference image.
`novelai_suggest_tags`	Danbooru/NovelAI tag autocomplete.

Generated images are saved to disk and also returned to the client, so the model can see them.

Related MCP server: Artifex MCP

Requirements

A NovelAI subscription and a Persistent API Token: NovelAI → Settings → Account → Persistent API Tokens.
uv. It runs the server without a manual install.

Cost

Requests spend your Anlas the same way the NovelAI website does. Large sizes, more steps, upscale and Vibe Transfer cost more. Check your Anlas balance before running many generations. The server does not limit spending.

Setup

Claude Code

claude mcp add novelai -e NOVELAI_API_KEY=your_token -- uvx mcp-server-novelai

Claude Desktop

Edit claude_desktop_config.json:

Windows: %APPDATA%\Claude\claude_desktop_config.json
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "novelai": {
      "command": "uvx",
      "args": ["mcp-server-novelai"],
      "env": {
        "NOVELAI_API_KEY": "your_token"
      }
    }
  }
}

Restart Claude Desktop after editing.

Cursor

Add the same mcpServers block to .cursor/mcp.json or to Settings → MCP.

Install from GitHub instead of PyPI

Replace "args": ["mcp-server-novelai"] with:

"args": ["--from", "git+https://github.com/DesmondFox/novelai-mcp", "mcp-server-novelai"]

Configuration

Variable	Required	Default	Meaning
`NOVELAI_API_KEY`	yes	—	Persistent API Token.
`NOVELAI_OUTPUT_DIR`	no	`~/Pictures/NovelAI`	Where images are saved. Each tool also accepts `output_dir`.

Usage

Ask the model in plain language. It picks the tool and parameters. Examples:

"Generate a portrait of a girl with silver hair and blue eyes in a sunflower field."
"Make a chibi cat girl holding a coffee mug, transparent background."
"Draw a tavern scene: a knight girl on the left, a mage boy on the right."
"Cyberpunk storefront at night with a neon sign that says NEON DREAMS."
"Take ~/Pictures/NovelAI/t2i_20260101_120000_abc123.png and change her expression to smug."
"Remove the background from that image, then upscale it."
"Suggest tags for 'kimono'."

Useful parameters of novelai_generate_image:

Parameter	Default	Notes
`model`	`nai-diffusion-5-full`	Also `nai-diffusion-5-curated`, `nai-diffusion-4-5-full`, `nai-diffusion-4-5-curated`.
`size`	`portrait`	`portrait` 832×1216, `landscape` 1216×832, `square` 1024×1024, `wallpaper` 1920×1088, `comic_strip` 1536×640, and others.
`width`, `height`	—	Custom size, rounded to a multiple of 64. Both must be set.
`steps`	23	Same default as the NovelAI website.
`scale`	5.0	Prompt guidance.
`seed`	random	Set it to reproduce an image. The used seed is returned.
`uc_preset`	`light`	`strong`, `light`, `furry_focus`, `human_focus`, `none`.
`characters`	—	List of `{"prompt": ..., "negative_prompt": ..., "position": [x, y]}`, x and y in 0.0–1.0. Up to 22 on V5, 6 on V4/V4.5.
`render_text`	—	Text to draw inside the image.
`transparent`	false	Transparent background.
`comic_panels`	false	Adds comic/manga panel tags.

Development

git clone https://github.com/DesmondFox/novelai-mcp
cd novelai-mcp
pip install -e ".[dev]"
pytest
ruff check .

For local development you can put the key in a .env file (see .env.example) instead of the client config.

License

MIT

Available Tools

7 tools

novelai_augmentB

NovelAI Director Tools for image manipulations: remove background ('bg-removal'), extract line art ('lineart'), convert to sketch ('sketch'), colorize sketch ('colorize'), change facial expressions/emotions ('emotion'), or declutter ('declutter').

Inputs: tool (required), image_path (required), defry, prompt, +2 more.

novelai_generate_imageB

Generate anime/art images with NovelAI Diffusion models (prioritizing V5 Full/Curated, V4.5). Supports natural language, Danbooru tags, multilingual prompts, transparent background, and custom dimensions.

Inputs: prompt (required), seed, size, model, +18 more.

novelai_img2imgA

Modify or restyle an existing image using NovelAI Image-to-Image. Applies new prompt with adjustable change strength (0.0 to 1.0) and optional noise.

Inputs: prompt (required), image_path (required), seed, size, +8 more.

novelai_inpaintB

Inpaint / redraw specific parts of an image using a mask. Allows replacing clothes, faces, hair, background or fixing anatomy.

Inputs: prompt (required), mask_path (required), image_path (required), seed, +7 more.

novelai_suggest_tagsB

Search and suggest valid Danbooru/NovelAI tags matching a query string.

Inputs: query (required), lang, model. Returns: result

novelai_upscaleB

Upscale an image with NovelAI's neural upscaler.

Inputs: image_path (required), model, output_dir.

novelai_vibe_transferC

Generate a new image influenced by the style, colors, and atmosphere of a reference image (Vibe Transfer).

Inputs: prompt (required), reference_image_path (required), seed, size, +8 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

B3.3/5.0

Scored 2026-09-22 20:50 across 7 tools

Disambiguation4/5

Each tool has a distinct purpose: upscaling, tag suggestion, generation, img2img, inpainting, augmentations, and vibe transfer. However, novelai_img2img and novelai_vibe_transfer both involve modifying or generating from an existing image, which could cause some confusion. The descriptions clarify the difference (strength vs style influence), but a user might initially hesitate.

Naming Consistency5/5

All tools follow a consistent 'novelai_' prefix followed by a clear verb or action (generate, img2img, inpaint, augment, vibe_transfer). The naming is uniform and predictable, making it easy to understand the tool's function from its name.

Tool Count5/5

7 tools is an appropriate scope for an image generation and editing MCP server. Each tool covers a distinct aspect of the workflow: generation, editing (img2img, inpaint), augmentation, style transfer, and utilities (upscale, tag suggestions). This is neither too thin nor too heavy.

Completeness4/5

The tool set covers the core image generation and editing workflows: generating new images, refining them, inpainting, style transfer, and enhancements. A notable gap is the lack of explicit image-to-image with mask (which is partially covered by inpaint) and perhaps batch processing, but the provided tools are sufficient for most common use cases. Minor gap: no explicit 'delete' or 'get' tools, but those are not typical for generation servers.

Maintenance

ActivityMaintained

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

InkRoomA
io.github.InkRoomAI
Generate, edit and upscale images with references, batch jobs, cost previews and original downloads.
AI & Machine Learning Image & Video Processing
121
https://mcp.inkroom.ai
Avocado AI
io.github.Wandersonjack
Generate AI images, videos, music, SFX & speech in any AI assistant. Results appear inline in chat.
AI & Machine Learning Image & Video Processing Marketing
2
https://mcp.avocadoai.co/api/mcp
mcpA
com.budgetpixel
Generate AI images, video, music, and sound effects, and upscale them, from any MCP client.
AI & Machine Learning Image & Video Processing Audio Processing
3113
https://mcp.budgetpixel.com/mcp
APICK AIA
app.apick
LLM chat, text tools, image generation, editing, batch image jobs, and asynchronous video generation
AI & Machine Learning Image & Video Processing
1511
https://apick.app/mcp/ai

Related MCP Servers

NovelAI MCP Server
AI & Machine Learning Image & Video Processing
hesoyamlp1
A
license
Not graded
quality
B
maintenance
Enables AI assistants to generate images using NovelAI, supporting text-to-image, image-to-image, and tag suggestions.
Updated 2 months ago (2026-09-15 12:09 UTC)
23 PyPI
3
MIT
Artifex MCP
AI & Machine Learning Image & Video Processing
jkalasas
A
license
A
quality
D
maintenance
Enables AI image generation via Antigravity (Google Gemini) and OpenAI DALL-E 3, supporting text-to-image, image editing, multiple outputs, and character consistency.
Updated 10 months ago (2026-01-24 03:07 UTC)
5
20 npm
9
MIT
openrouter-imgen-mcp
AI & Machine Learning Image & Video Processing
jtxmp
A
license
A
quality
D
maintenance
Gives AI assistants image generation and editing capabilities through OpenRouter, supporting multiple models, style presets, variations, and batch operations.
Updated 8 months ago (2026-03-07 12:17 UTC)
5
75 npm
2
MIT
MCP Image Generator
AI & Machine Learning Image & Video Processing
Supachai812
A
license
Not graded
quality
A
maintenance
Enables assistants in any MCP client to generate new images or edit existing ones, automatically enriching short requests with composition, lighting, and camera details before sending them to Gemini, OpenAI, or BytePlus Seedream. Supports quality presets, aspect ratios, image sizes, character consistency, blended inputs, and Google Search grounding for current information.
Updated a month ago (2026-10-17 21:35 UTC)
1,781 npm
MIT

NovelAI MCP Server

Tools

Requirements

Cost

Setup

Claude Code

Claude Desktop

Cursor

Install from GitHub instead of PyPI

Configuration

Usage

Development

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

NovelAI MCP Server

Artifex MCP

openrouter-imgen-mcp

MCP Image Generator