Connect BlendCanvas to your AI client
Connect BlendCanvas to Claude Code, Codex, Cursor, GitHub Copilot, and other tools that support remote MCP.
Set up your connection
Remote MCP is currently unavailable for this workspace.
Let AI set up the connection
Paste the request into your AI tool. You can choose your tool to include its exact setup instructions.
Help me add or update the blendcanvas connection in this client’s user settings using Streamable HTTP at https://app.blendcanvas.com/api/mcp. Preserve my other MCP connections and approval settings. Guide me to create a token at https://app.blendcanvas.com/account?tab=mcp&lang=en#mcp-connection and enter it locally as a Bearer token. Never ask for the token in chat or print it. After I authorize the connection, verify that list_canvases is available. If you cannot edit the settings or I need to restart the client, tell me exactly what to do.
Manual setup and commands
Choose your tool and add the BlendCanvas entry to your existing user settings.
Claude Code
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
claude mcp add-json --scope user blendcanvas '{"type":"http","url":"https://app.blendcanvas.com/api/mcp","headers":{"Authorization":"Bearer ${BLENDCANVAS_MCP_TOKEN}"}}'Codex
~/.codex/config.toml
Replace YOUR_MCP_TOKEN in your private client settings.
[mcp_servers.blendcanvas]
url = "https://app.blendcanvas.com/api/mcp"
http_headers = { Authorization = "Bearer YOUR_MCP_TOKEN" }Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
codex mcp add blendcanvas --url https://app.blendcanvas.com/api/mcp --bearer-token-env-var BLENDCANVAS_MCP_TOKENCursor
~/.cursor/mcp.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}GitHub Copilot (VS Code / CLI)
~/.copilot/mcp-config.json
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
{
"mcpServers": {
"blendcanvas": {
"type": "http",
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer ${BLENDCANVAS_MCP_TOKEN}"
},
"tools": [
"*"
]
}
}
}Windsurf
~/.codeium/windsurf/mcp_config.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"serverUrl": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}TRAE
TRAE → Settings → MCP → Raw Configuration (JSON)
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Google Antigravity
~/.gemini/config/mcp_config.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"serverUrl": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Cline
Cline → MCP Servers → Configure MCP Servers
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"type": "streamableHttp",
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Kiro
~/.kiro/settings/mcp.json
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
{
"mcpServers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer ${BLENDCANVAS_MCP_TOKEN}"
}
}
}
}Amazon Q Developer
Amazon Q → Chat → Tools → Add MCP server → Global
Replace YOUR_MCP_TOKEN in your private client settings.
Name: blendcanvas
Transport: http
URL: https://app.blendcanvas.com/api/mcp
Authorization: Bearer YOUR_MCP_TOKENJunie (JetBrains)
~/.junie/mcp/mcp.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}JetBrains AI Assistant
Settings → Tools → AI Assistant → Model Context Protocol → HTTP → Global
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"type": "streamable-http",
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Zed
zed: open settings file
Replace YOUR_MCP_TOKEN in your private client settings.
{
"context_servers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Gemini CLI
~/.gemini/settings.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"httpUrl": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}Qwen Code
~/.qwen/settings.json
Replace YOUR_MCP_TOKEN in your private client settings.
{
"mcpServers": {
"blendcanvas": {
"httpUrl": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_MCP_TOKEN"
}
}
}
}OpenCode (v2)
~/.config/opencode/opencode.json
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
{
"mcp": {
"servers": {
"blendcanvas": {
"type": "remote",
"url": "https://app.blendcanvas.com/api/mcp",
"oauth": false,
"headers": {
"Authorization": "Bearer {env:BLENDCANVAS_MCP_TOKEN}"
}
}
}
}
}Kilo Code
~/.config/kilo/kilo.jsonc
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
{
"mcp": {
"blendcanvas": {
"type": "remote",
"url": "https://app.blendcanvas.com/api/mcp",
"oauth": false,
"headers": {
"Authorization": "Bearer {env:BLENDCANVAS_MCP_TOKEN}"
}
}
}
}Continue
~/.continue/config.yaml
Replace YOUR_MCP_TOKEN in your private client settings.
mcpServers:
- name: blendcanvas
type: streamable-http
url: "https://app.blendcanvas.com/api/mcp"
requestOptions:
headers:
Authorization: "Bearer YOUR_MCP_TOKEN"Amp
~/.config/amp/settings.json
Set BLENDCANVAS_MCP_TOKEN in the environment that starts your client, then use the configuration below.
{
"amp.mcpServers": {
"blendcanvas": {
"url": "https://app.blendcanvas.com/api/mcp",
"headers": {
"Authorization": "Bearer ${BLENDCANVAS_MCP_TOKEN}"
}
}
}
}Try a canvas request
After connecting, ask your AI client to run this request.
Use the BlendCanvas MCP tool list_canvases to list my canvases. Then use get_canvas_summary to summarize the first canvas. If I have no canvases, say so.
Hosted canvas tools
- List my canvases
list_canvases - Find canvases in your BlendCanvas account.
- Read the canvas
get_canvas_summary - Read the title, nodes, and connections of a canvas you own.
- Create a canvas
create_canvas - Create an empty canvas in your BlendCanvas account.
- Draft on the canvas
create_text_node - Add a text note to a canvas you own.
Reload the workspace to see canvas changes made by your client.
Available assistant skills
Open workspaceChoose a system skill in the canvas assistant to guide a project. Skill names and descriptions are shown as published.
H3 3D动画短片生成器
将一句故事创意转化为完整的风格化3D动画短片制作流程,依次完成项目简报、故事大纲、角色卡、无人物场景卡、标准镜头表、文本分镜或可选铅笔分镜、当前视频模型确认、单镜头生成、片段顺序编组、BGM匹配方案与最终复查。适用于需要角色一致、场景连续、节奏明确和音画配合的剧情动画短片、纪念短片、品牌故事与社交媒体内容。不适用于单张图片、简单修图、真人写实视频或单一独立镜头。
H3 品牌宣传视频生成器
面向需要为品牌、产品、网站、App、小店或个人项目制作宣传内容的运营与创作者。用户需提供 LOGO、产品图、界面截图、官网链接或其他可核验素材,并确认时长、画幅、受众和推广重点。Skill 会整理品牌事实与素材来源,选择叙事方向,规划精确节拍和镜头,并使用当前画布已有的图像、视频、音频、文本和分组能力生成宣传资产。适用于新品发布、官网展示和社交媒体推广,不适用于缺少授权素材时仿造真实品牌标识、虚构产品功能或制作长篇剧情影片
H3 合作游戏开场动画生成器
面向想制作双人合作游戏主菜单或开场动画的用户。用户需提供两位玩家名称、游戏名称、目标视觉风格,并可上传角色参考图。Skill 会先锁定人物身份特征,基于固定菜单布局生成与色彩、按钮、图标和字体联动的确认首图;用户确认后,回填角色、界面文案和事件节奏,生成完整视频提示词并制作最终开场视频。最终输出一条包含双人角色、玩家信息卡和主菜单交互动效的游戏开场视频。适用于合作游戏概念展示、角色化菜单和社交内容,不适用于完整可玩游戏开发、复杂多页面 UI、精确品牌标识复刻或无角色的通用片头
H3 音乐视频字幕生成器
面向音乐人、视频创作者和社交媒体剪辑者,用于制作带歌词贴字的 AI MV 或情绪短片。用户提供音乐、歌词、参考图、角色、字体方向、情绪意图或发布平台。Skill 会依据用户提供或确认的节拍、BPM、歌词和人声时间点,区分人物、场景和文字参考,设计随节奏变化的空间字幕,拆解长视频为可衔接镜头,审查提示词并调用当前全局视频模型。最终输出 MV 概念、分镜提示词、歌词文字方案和拼接建议。适用于风格化音乐视觉和动态字幕 MV,不适用于普通字幕校对、照搬已有 IP 或完全手工后期剪辑。
H3 剪纸定格动画
面向内容创作者、教育讲解者和社交视频编辑,用触感纸拼贴语言表现口播句、知识点、观点或抽象主题。用户提供短文案、故事节点或核心概念,也可补充画幅、时长、色调和音频需求。Skill 会提炼含义与视觉隐喻,制定制作方案和分镜,生成并确认半调纸拼贴静帧,再制作带纸片滑入、弹入、轻敲、压平和摩擦声的停格动画片段;多片段任务按顺序编组并提供拼接说明。默认保留拼贴音效,不主动添加 BGM、旁白和字幕。适合知识讲解、观点表达、故事配画和社交 B-roll,不适合真人口播广告、精确可编辑图层、复杂文字排版或仅输出提示词
H3 极简产品广告生成器
上传一张产品图,就能按 Apple 风做一条高级产品广告短片。适合电商商品、小品牌新品、潮玩、食品、数码配件等产品展示:我会先帮你确认时长、比例和风格,再提炼产品卖点、生成英文广告文案、制作产品锚定图、规划文字卡点分镜,并生成一条干净、有节奏、产品质感突出的短片。不适合 KOC 口播、普通剪辑或复杂屏幕演示
H3 官方默认 Skill
为 MiniMax H3 的五种视频生成模式(T2VA、I2VA、FL2VA、L2VA 和 Ref2VA)编写结构化的提示信息
Open the workspace to see the current assistant skills.
The hosted workspace includes a canvas assistant powered by managed cloud models. Sign in to describe an idea and continue working on the canvas.