Aidelly MCP Server
Automate Aidelly social media management directly from Claude Code. Create, schedule, and manage posts across all connected platforms with AI-powered MCP tools.
README
Aidelly Claude Code Plugin
Automate Aidelly social media management directly from Claude Code. Create, schedule, and manage posts across all connected platforms with AI-powered MCP tools.
Features
- Create & Schedule Posts – Publish instantly or schedule for later across multiple platforms
- Approval Workflows – Require and manage post approvals before publishing
- Analytics – Pull engagement metrics and top-performing content
- Multi-Platform Support – Instagram, Facebook, LinkedIn, Twitter/X, TikTok, YouTube, Pinterest, Bluesky, Threads, Mastodon, Google Business
- Media Upload – Fetch and upload images/videos directly
Installation
Sign-in is one click. The Aidelly MCP server supports OAuth 2.1 with PKCE and dynamic client registration, so Claude walks you through an Aidelly login and consent screen automatically — no API key or workspace ID to copy. (An API-key header remains supported for scripted/headless use.)
Install via Claude Code Plugin Marketplace
/plugin marketplace add Aidelly/claude-plugin
Then configure your API key (see Setup below).
Manual Installation (Development)
Clone this repository and point Claude Code to the local plugin path:
git clone https://github.com/Aidelly/claude-plugin
cd claude-plugin
# Add to Claude Code plugin settings
Note: Ensure
.mcp.jsonis committed in your local clone. The Aidelly monorepo ignores*.mcp.jsonfiles, so the publicAidelly/claude-pluginrepository explicitly commits it. If you see a missing.mcp.jsonafter cloning, rungit check-ignore .mcp.json— if no output, it's been properly committed.
Setup
1. Generate an API Key
- Log in to Aidelly Dashboard
- Go to Settings → API Keys
- Create a new API key (or use an existing one)
- Copy the key
2. Configure Environment Variables
Set the following environment variables in your Claude Code environment:
export AIDELLY_API_KEY=your_api_key_here
export AIDELLY_WORKSPACE_ID=your_workspace_id # Optional; required if managing multiple workspaces
macOS/Linux (shell):
# Add to ~/.zshrc, ~/.bashrc, or equivalent
echo 'export AIDELLY_API_KEY=sk_...' >> ~/.zshrc
source ~/.zshrc
Windows (PowerShell):
[Environment]::SetEnvironmentVariable('AIDELLY_API_KEY', 'sk_...', 'User')
3. Verify Connection
In Claude Code, run:
Use the MCP server to list accounts or get workspace info.
The plugin should successfully connect and list your Aidelly workspace data.
Usage
Create a Post
Create a new post on Instagram with the text "Hello world!" and upload media.
Schedule a Post
Schedule a post for tomorrow at 2 PM with approval required.
Check Pending Approvals
List all posts waiting for approval and approve the first one.
Pull Analytics
Get the top 10 posts from the last 30 days and their engagement metrics.
Skill Documentation
Full endpoint reference and examples: skills/aidelly-social/SKILL.md
Configuration
.mcp.json
The plugin wires to Aidelly's MCP server via .mcp.json:
{
"mcpServers": {
"aidelly": {
"url": "https://app.aidelly.ai/api/mcp/public-api",
"transportType": "http",
"auth": {
"type": "headers",
"headers": {
"Authorization": "Bearer ${AIDELLY_API_KEY}",
"x-aidelly-workspace-id": "${AIDELLY_WORKSPACE_ID}"
}
}
}
}
}
plugin.json
Metadata for Claude Code plugin marketplace:
{
"name": "aidelly",
"description": "Aidelly social content automation for Claude Code",
"author": "Aidelly",
"version": "0.1.0"
}
Troubleshooting
"Unauthorized" Error
- Verify
AIDELLY_API_KEYis set correctly - Confirm the API key is active in Aidelly Dashboard
- Check that the workspace exists
"Forbidden" Error
- Ensure your API key has access to the workspace
- If using
AIDELLY_WORKSPACE_ID, verify it matches a workspace you have access to
Rate Limit Exceeded
- The plugin enforces 120 tool calls per minute per API key
- Retry after a few seconds with exponential backoff
Platform-Specific Issues
- Instagram/TikTok deletion: These platforms don't allow post deletion via API
- LinkedIn org routing: Double-check that the author target matches your account type (member vs. organization)
- Text too long: Verify text length against platform limits in the skill documentation
Architecture
.claude-plugin/
└─ plugin.json # Plugin metadata
.mcp.json # MCP server configuration
skills/
└─ aidelly-social/
└─ SKILL.md # Tool documentation & examples
API Endpoints
The plugin connects to https://app.aidelly.ai/api/mcp/public-api (Aidelly's MCP HTTP endpoint).
Authentication: Authorization: Bearer <API_KEY> (required)
Tools exposed by the server:
aidelly_create_post– Instant post publishaidelly_create_scheduled_post– Schedule a postaidelly_list_pending_approvals– Check pending approvalsaidelly_post_approvals_id_action– Approve/reject postsaidelly_get_analytics_summary– Pull engagement metricsaidelly_get_post– Get post detailsaidelly_upload_media– Upload image/video- Full list of available tools
Publishing & Distribution
This repository is designed to be published to Claude Code plugin marketplaces. Follow the checklist below for each distribution channel.
Distribution Channels
1. Anthropic Claude Code Plugin Directory
Status: [Pending / In progress / Published]
Submission checklist:
- [ ] README.md complete with features, installation, setup
- [ ] .claude-plugin/plugin.json valid and descriptive
- [ ] .mcp.json correctly configured for production MCP endpoint
- [ ] Skills documentation comprehensive with examples
- [ ] All JSON files validate (no parse errors)
- [ ] Version bumped to match release (semver)
- [ ] LICENSE file included (if required)
- [ ] Documentation reviewed for accuracy
How to submit:
- Contact Anthropic at plugins@anthropic.com with:
- GitHub repository link
- Brief description of plugin capabilities
- API authentication requirements
- Follow Anthropic's submission process
- Monitor distribution portal for approval status
2. Official MCP Registry (Smithery)
Status: [Pending / In progress / Published]
Registry: https://smithery.ai or https://registry.smithery.ai
Submission checklist:
- [ ] Repository public on GitHub
- [ ] .mcp.json follows MCP spec v1.0+
- [ ] Version in package.json matches .claude-plugin/plugin.json
- [ ] README includes authentication setup
- [ ] Endpoint URL is production-ready (https://)
- [ ] Rate limits documented
- [ ] Error handling documented
- [ ] Test credentials available for verification (optional)
How to submit:
- Fork the MCP registry (https://github.com/smithery/smithery-registry)
- Add entry in registries.json (or equivalent)
- Include metadata: name, version, description, auth requirements, URL
- Open pull request
- Address review feedback
3. Alternative MCP Registries
mcp.so – https://mcp.so PulseMCP – https://pulsemcp.com Glama – https://glama.ai
Submission checklist (each):
- [ ] Plugin discoverable via their search
- [ ] Correct metadata (name, description, author)
- [ ] Production MCP endpoint working
- [ ] Authentication method documented
- [ ] Contact/support information provided
How to submit:
- Each registry has its own submission process (check their websites)
- Generally requires GitHub repo + metadata + contact info
Pre-Release Checklist
Before submitting to any marketplace:
# 1. Validate JSON files
jq . .claude-plugin/plugin.json
jq . .mcp.json
# 2. Verify MCP endpoint connectivity
curl -H "Authorization: Bearer $AIDELLY_API_KEY" \
https://app.aidelly.ai/api/mcp/public-api \
-X POST \
-d '{"jsonrpc":"2.0","method":"initialize","id":1}' \
-H "Content-Type: application/json"
# 3. Test API key setup
export AIDELLY_API_KEY=test_key
echo "API key set: $AIDELLY_API_KEY"
# 4. Review skill documentation
cat skills/aidelly-social/SKILL.md | wc -l # Should be > 200 lines
Release Notes
When bumping version in .claude-plugin/plugin.json and README:
## Version 0.1.0
**Initial release**
- Create, schedule, and publish posts across multiple platforms
- Approval workflow integration
- Analytics summary pull
- Media upload support
- 11 platforms supported
Support
- Issues: https://github.com/Aidelly/claude-plugin/issues
- Docs: https://docs.aidelly.ai
- Email: support@aidelly.ai
License
[License TBD – check Aidelly open-source policy]
Maintainers
- Aidelly Team (@aidelly)
Last updated: 2026-07-13
Plugin version: 0.1.0
MCP protocol version: 2024-11-05+
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。