Krita Illustration MCP
Provides tools for controlling a local Krita instance, including creating canvases and layers, importing PNGs, defining and applying masks, inpainting with generated images, painting with Krita's brush engine, drawing primitives, checkpointing KRA files, previewing the canvas, and saving/exporting documents.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Krita Illustration MCPCreate a new transparent canvas and import lineart.png as a separate layer"
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.
Krita Illustration MCP
让 AI 大模型通过 MCP 直接控制本机 Krita。模型查看 krita_preview 返回的真实画布,决定下一步文档、图层和绘制操作;Krita Bridge 在主线程执行,所以变化会直接显示在 Krita 中。
复杂人物插画优先使用混合流程:客户端生成参考图 → 客户端生图工具转换线稿 → MCP 导入独立图层 → Krita 局部修正和导出。生成与原生笔画必须明确区分,不把导入 PNG 冒充逐笔绘制或自动语义分层。服务本身默认不调用外部生图 API、不读取厂商密钥;GPT/Gemini 仍只保留未配置的扩展接口。
当前能力
创建透明画布和独立图层。
krita_image检查 PNG 尺寸/哈希,并以来源标记导入新图层;保留图片副本,重复请求不重复导入。查询文档/图层状态并返回真实 PNG 预览。
定义、裁剪、预览区域并创建/扩展/羽化蒙版。
用
krita_inpaint将客户端或已配置 provider 生成的局部 PNG 按 Mask 应用为新图层。创建 KRA checkpoint,恢复为新文档。
安全保存 KRA 和导出 PNG。
对关键原生修改使用持久化
request_id去重。用
krita_paint调用 Krita Brush Engine 批量绘制有压力信息的笔画。用
krita_draw绘制直线、折线、贝塞尔曲线、矩形、椭圆和可填充多边形。用
krita_illustration持久化草稿、线稿、底色、阴影、高光、终检和导出阶段,未预览不能推进。
原生图元、局部清除、预览和保存已经通过牡丹绘制实机验证;人物逐曲线重画成本较高,因此不再作为复杂插画的推荐主流程。图片入口复用现有 Bridge import_image,本次仅需重新加载 MCP 服务,不必重新安装 Bridge。
Related MCP server: GIMP MCP
目录
src/krita_workbench/
domain/ 纯数据与规则
application/ Krita 原生控制用例
contracts/ MCP 与 Bridge 共享契约
adapters/krita/ 本机 Bridge 客户端
adapters/storage/ job、region、checkpoint 持久化
native/ Krita 插件与主线程操作
transport/mcp.py STDIO MCP 工具
docs/ 架构、契约、开发计划、启动说明
dist/ Krita 插件 ZIP安装 Krita Bridge
在 Krita 中选择“工具 → 脚本 → 导入 Python 插件”,导入:
D:/krita-mcp/dist/krita_workbench_bridge-20260905.zip启用插件并重启 Krita。详细步骤见 native-bridge.md。
使用 uv 启动 MCP
先确保 D:/krita-mcp/config.local.json 与其中引用的 Bridge 配置存在,然后在 PowerShell 执行:
cd D:/krita-mcp
uv run python -m krita_workbench --config D:/krita-mcp/config.local.json --check-config
uv run python -m krita_workbench --config D:/krita-mcp/config.local.json第二条命令启动 STDIO MCP 服务,等待 MCP 客户端连接;它不是交互式命令行。
可使用 config/mcp.example.json 配置客户端直接启动该服务。
推荐调用顺序
复杂插画采用以下流程,参数示例见 混合绘图流程:
客户端生图并确认参考图,再以参考图为编辑输入生成干净线稿。
将 PNG 复制到
D:/krita-mcp/artifacts/,不扩大 Bridge 的 allowed roots。krita_image(action="inspect")获取真实尺寸,再新建匹配画布。分别导入 reference 和 lineart;隐藏 reference,预览检查线稿。
对手、眼睛、头发或褶皱定义 region,创建/扩展/羽化 Mask。
用客户端生成与 region 尺寸完全一致的修复 PNG,调用
krita_inpaint;结果会按 Mask 变成透明补丁并新增图层。预览修复结果,在新的修正图层用
krita_paint/krita_draw补线;危险修改前 checkpoint。最终预览后
krita_save_export;导入的是栅格层,不是自动拆出的头发/皮肤/服装层。
需要全程原生绘制时,原有流程仍可选:
krita_capabilitieskrita_statekrita_document创建透明画布和目标图层krita_illustration(action="start")记录文档和五个目标图层通过
krita_paint/krita_draw分批写入当前阶段图层每个主要阶段后
krita_preview,再用krita_illustration(action="advance")记录通过高风险修改前
krita_history(action="checkpoint")完成后
krita_save_export,记录 export 阶段并完成会话
krita_capabilities.external_image_generation 会显示保留的 gpt、gemini 槽位。默认 enabled=false、configured=false;只有未来显式实现并注入 provider adapter 后才会启用。
验证
测试只使用 mocks 和临时目录,不访问真实 Krita:
$env:PYTHONPATH='D:/krita-mcp/src;D:/krita-mcp'
D:/krita-mcp/.venv/Scripts/python.exe -m unittest discover -s tests -t . -v离线测试通过不等于真实 Krita 绘制验收。每个新增原生能力必须另做运行中 Krita 的可见操作和 preview 检查。
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Remote MCP for RunComfy: ComfyUI deployments, hosted models, LoRA training. 31 tools.
Build and run visual creative-production workflows from your AI agent.
Generate AI images and videos from any compatible MCP client.
Control Unreal Engine to browse assets, import content, and manage levels and sequences. Automate…
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables control of Open Brush 3D painting software through natural language commands, allowing LLMs to create drawings, manipulate brushes, import 3D models, manage colors, layers, and camera settings via the Open Brush HTTP API.MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI models and external applications to control GIMP remotely via the Model Context Protocol, allowing image manipulation and object querying through natural language.3MIT
- AlicenseCqualityBmaintenanceEnables AI assistants to control Aseprite for drawing and editing images through its API.59505MIT
- AlicenseNot gradedqualityCmaintenanceGive AI agents full control over your local ComfyUI by exposing 77 tools for workflow management, image generation, model handling, and real-time canvas control.3AGPL 3.0
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/xqqcylh6780/krita-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server