# AdaptlyPost Agent

[![smithery badge](https://smithery.ai/badge/tarasshyn/adaptlypost)](https://smithery.ai/servers/tarasshyn/adaptlypost)

Give your AI agent the ability to post to Instagram, TikTok, YouTube, X, LinkedIn, Facebook, Pinterest, Threads, and Bluesky from a single command, and to read how those posts performed.

**Supports:** Instagram, TikTok, YouTube, X (Twitter), LinkedIn, Facebook, Pinterest, Threads, Bluesky

## Install

```bash
npx skills add adaptlypost/agent
```

Works with Claude Code, Cursor, Windsurf, Codex, and any agent that supports skills.

### Other installation methods

**Manual**: Copy the `skills/adaptlypost/` folder into your project's skills directory.

**Cursor remote rules**: Point to `https://raw.githubusercontent.com/adaptlypost/agent/main/skills/adaptlypost/SKILL.md`

## Setup

1. Create an account at [adaptlypost.com](https://adaptlypost.com)
2. Connect your social media accounts
3. Create an API token at [Settings > API Tokens](https://adaptlypost.com/api-tokens)
4. Run:
   ```bash
   ./scripts/adaptlypost.js setup --key adaptly_xxxxx
   ```

## What it does

Once installed, your AI agent can:

- **Post** to 9 platforms simultaneously with per-platform caption overrides
- **Schedule** posts for any future time
- **Bulk schedule** up to 100 posts at once
- **Check results** per-platform success/failure with error details
- **Retry** just the failed platforms
- **Draft & publish** workflow — save drafts, review, publish later
- **Read analytics** — views, likes, comments, followers and engagement per window, per platform and per post

## Example

```
You: Schedule a LinkedIn post for tomorrow at 9am about our new feature launch
Agent: Done. Post scheduled for tomorrow at 9:00 AM on your LinkedIn account.
```

## Alternative: MCP

For deeper integration with Claude Desktop, Cursor, or other MCP-compatible clients, use the AdaptlyPost MCP server:

```json
{
  "mcpServers": {
    "adaptlypost": {
      "type": "http",
      "url": "https://mcp.adaptlypost.com/mcp",
      "headers": {
        "Authorization": "Bearer adaptly_your_key"
      }
    }
  }
}
```

[MCP setup docs](https://adaptlypost.com/features/agents)

### Run it locally

The server source lives in [`mcp-server/`](./mcp-server). Run it over stdio with [Bun](https://bun.sh):

```json
{
  "mcpServers": {
    "adaptlypost": {
      "command": "bun",
      "args": ["run", "/path/to/agent/mcp-server/src/index.ts"],
      "env": { "ADAPTLYPOST_API_TOKEN": "adaptly_your_key" }
    }
  }
}
```

Or with Docker:

```bash
docker build -t adaptlypost-mcp .
docker run -i -e ADAPTLYPOST_API_TOKEN=adaptly_your_key adaptlypost-mcp
```

## Links

- [AdaptlyPost](https://adaptlypost.com)
- [AI Agents page](https://adaptlypost.com/features/agents)
- [API Tokens](https://adaptlypost.com/api-tokens)
- [MCP Server](./mcp-server)

## License

MIT