Install Instagram MCP Server

Pick your client, copy the command, done.

claude mcp add -e "INSTAGRAM_ACCESS_TOKEN=${INSTAGRAM_ACCESS_TOKEN}" instagram-mcp-7b6f -- python path/to/ig-mcp/main.py

Required:INSTAGRAM_ACCESS_TOKEN

Environment Variables

Set these before running Instagram MCP Server.

VariableDescriptionRequired

INSTAGRAM_ACCESS_TOKENLong-lived access token with appropriate permissionsYes

Once configured, Instagram MCP Server gives your AI agent access to:

get_profile_infoRetrieve Instagram business profile details

get_media_postsFetch recent posts from an Instagram account

get_media_insightsRetrieve engagement metrics for specific posts

post_id

publish_mediaUpload and publish images/videos to Instagram

image_urlcaption

get_account_pagesList Facebook pages connected to the account

get_conversationsList Instagram DM conversations

get_conversation_messagesRead messages from specific conversations

conversation_id

send_dmReply to Instagram direct messages

conversation_idmessage

After setup, try these prompts with your AI agent:

→Analyze the engagement metrics for my most recent Instagram post.

→Draft a new Instagram post about our latest product launch and publish it.

→List my recent Instagram DM conversations and summarize the pending inquiries.

→What is the current follower count and bio for my business profile?

→Find the top performing post from last week based on engagement insights.

Prerequisites & system requirements

An MCP-compatible client (Claude Code, Cursor, Windsurf, Claude Desktop, or Codex)
Python 3.8+ with pip installed
INSTAGRAM_ACCESS_TOKEN — Long-lived access token with appropriate permissions

Save the docs, env vars, and workflow around Instagram MCP Server in Conare so Claude Code, Codex, and Cursor remember it next time.