The social media API
built for AI agents.
Let your AI post to Instagram, TikTok, YouTube, X, LinkedIn, Facebook, Threads, Pinterest, and Bluesky. Works with Claude, ChatGPT, Cursor, or any agent you build.
Free trial · No credit card · Works with any MCP-aware agent
Connect your AI agent to social media.
Post to 9 platforms — from any AI agent, any framework, one API.
MCP
Claude, ChatGPT, Cursor, Windsurf
Native Model Context Protocol integration. Your AI assistant can schedule posts, check analytics, and manage accounts — directly from chat.
REST API
Custom agents & workflows
Build your own integration with a clean REST API. Endpoints for posting, scheduling, media uploads, and analytics. Any language, any framework.
Autonomous agents
Let your bot run the calendar
Give your background agent a key and let it plan, write, schedule, and analyze — without you in the loop. Same auth as MCP.
Webhooks (coming)
Push events to your stack
Subscribe to post.published, post.failed, dm.received, and more. We'll POST to your URL with a signed payload — let your agent react in real time.
Everything your agent needs.
7 tools via MCP. Full REST API. Your AI gets complete control.
Multi-platform posting
Post to 9 platforms simultaneously. Per-platform caption overrides, media, and scheduling.
Schedule & queue
Schedule posts for any time. Your agent manages the calendar so you don't have to.
Analytics on demand
Ask your AI for views, likes, comments, and shares. Aggregated across every platform.
Media uploads
Upload images and videos via file or URL. Auto-stored in your media library.
Account management
List connected accounts, check status, and route posts to the right ones.
Per-platform error handling
If one platform fails, the others still go through. Your agent can retry or alert you.
One config block.
Drop this into your Claude Desktop, Cursor, or Windsurf MCP config. Your agent gets 7 tools instantly.
{
"mcpServers": {
"flowtive": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/inspector",
"https://flowtive.app/api/mcp"],
"env": {
"FLOWTIVE_API_KEY": "fl_live_..."
}
}
}
}POST https://flowtive.app/api/mcp
Authorization: Bearer fl_live_...
Content-Type: application/json
{ "jsonrpc": "2.0", "id": 1, "method": "tools/list" }Get a key in Settings → Developer. The same key works for MCP and the REST API.
One endpoint, every platform.
A single POST creates a post on every connected account, with optional per-platform overrides.
curl -X POST https://flowtive.app/api/v1/posts \
-H "Authorization: Bearer fl_live_..." \
-H "Content-Type: application/json" \
-d '{
"caption": "Just shipped a new feature. Here is what it does ↓",
"platforms": ["twitter", "linkedin", "threads", "bluesky"],
"scheduled_at": "2026-06-15T14:00:00Z"
}'curl https://flowtive.app/api/v1/analytics \
-H "Authorization: Bearer fl_live_..."
# {
# "totals": { "views": 12483, "likes": 421, "comments": 38, ... },
# "posts": [ ... ]
# }curl https://flowtive.app/api/v1/accounts \
-H "Authorization: Bearer fl_live_..."How agents actually use this.
From shipping announcements to fully autonomous content bots.
Ship announcements from your terminal
Pushed a new feature? Tell your AI to announce it on every platform. No context switching, no opening dashboards.
Run a fully autonomous content agent
Build an AI agent that researches trends, writes posts, schedules them, and reports back — 24/7 without your input.
Manage multiple brands or clients
Agencies: connect every client account once, give each a workspace, and let agents post to the right one with the right voice.
Wire posting into your CI/CD
Auto-post release notes, stats, or build artifacts. The API fits anywhere your code runs.
Common questions.
Do I need to install anything?▼
For HTTP MCP: nothing — drop the URL into your client's config and you're done. For Claude Desktop's local mode you'll use the standard MCP inspector wrapper. For the REST API: just curl it.
What platforms can my agent post to?▼
Instagram, TikTok, YouTube, X (Twitter), LinkedIn, Facebook, Pinterest, Threads, and Bluesky. All 9 from a single API call — connect each in the dashboard once, then your agent picks any of them per post.
What's the difference between MCP and REST?▼
MCP is a protocol — AI assistants like Claude use it natively, so the agent gets typed tool calls instead of free-form prompts. REST is a standard HTTP API for custom integrations. Both share the same backend and the same API key.
Can my agent upload images and videos?▼
Yes. POST to /api/v1/media with a file (multipart) or a public URL. We'll host it, return a usable URL, and your agent passes it to /api/v1/posts.
Can I schedule posts for later?▼
Yes — pass a scheduled_at ISO timestamp. Our scheduled-posts cron picks it up and publishes at the right time, even if your agent has gone offline.
What if a post fails on one platform?▼
Each platform posts independently. The response tells you per-platform: which succeeded, which failed, and why. Your agent can retry the failures or alert you.
Is there a rate limit?▼
60 requests/minute per key with burst up to 30. More than enough for normal agent traffic. Hit the ceiling and you'll get a 429 with a Retry-After header.
Give your AI agent social media.
MCP for AI assistants. REST for everything else. Same backend, same key.