Instagram MCP server for your AI assistant

What your agent can pull from Instagram

Ask in everyday words — the agent fetches the media, the account that posted it and the audience around it

• Instagram profile

  Bio, counts of followers and posts, verified and business flags.

• Account search

  Look an account up by name when the exact username slips your mind.

• Posts and reels

  Recent posts and reels from an account, captions and media included.

• Post by URL

  A single post, reel or carousel — all its slides, the caption and the author.

• Comments

  Everything written under a post or reel, with authors and dates.

• Likers

  Accounts that liked a post or reel, with their names and avatars.

• Followers

  The follower list of an account, page by page.

• Following

  The accounts it follows in return, page by page.

• Stories live now

  Whatever an account has live at this moment, before it expires.

• Story by URL

  One story: its media, the author and the time it disappears.

• Highlights

  Saved collections together with every story kept inside them.

• Video download

  Direct links to the video behind a post, reel or IGTV — without a watermark.

A single MCP for the whole of Instagram

One set of tools for a chat client, an editor or your own code — reels, photos, videos, carousels, stories and highlights.

{
  "mcpServers": {
    "igram-mcp": {
      "type": "http",
      "url": "https://api-wh.igram.world/instagram/mcp"
    }
  }
}

Connecting the Instagram MCP to your client

Choose your platform and follow the steps

Universal setup Universal ChatGPT ChatGPT Claude Claude

🔌 Any client that speaks MCP will do · only the path to the config file changes.

1. Open the MCP settings of your client and locate its config file — usually mcp.json.
2. Add iGram inside the mcpServers block:
   mcp.json Copy

   {
     "mcpServers": {
       "igram-mcp": {
         "type": "http",
         "url": "https://api-wh.igram.world/instagram/mcp"
       }
     }
   }

3. Save the file and restart the client so the server is picked up.
4. Now ask for something in a chat:
   • Grab the video from this post: instagram.com/p/…
   • Save every story this account has live right now

One key, every tool included

• Pro

  $50.00 /mo

  What you get:

  • 50,000 requests per month
  • A hard cap — no surprise overage
  • Proxies included
  • MCP access
  Try it for free
• Best value

  Ultra

  $150.00 /mo

  What you get:

  • 250,000 requests per month
  • A hard cap — no surprise overage
  • Proxies included
  • MCP access
  Try it for free

• Mega

  $250.00 /mo

  What you get:

  • 1,000,000 requests per month
  • A hard cap — no surprise overage
  • Proxies included
  • MCP access
  Try it for free

Ask for a payment link in the same chat where you connected iGram — ChatGPT, Claude or any other client — and your assistant will hand one over.

Need more than that? Write to us about a custom plan.

Questions we get asked

Everything about wiring iGram into an AI agent.

• What exactly is MCP?

  The Model Context Protocol is an open standard that wires AI assistants to outside tools and data. A server publishes a set of tools, and the assistant works out when to call them.

• Which clients are supported?

  Any client that speaks MCP. The setup section above walks through ChatGPT and Claude — on the web, in the desktop app and in Claude Code — and a universal config covers the rest.

• What about private accounts?

  They are out of reach. Only publicly available content can be fetched, just as on the website.

• How does this differ from the REST API?

  The data is the same, the consumer is not. The REST API serves code you write yourself; MCP serves an agent that judges on its own when something needs downloading.