Instagram MCP Server: AI Agent for Instagram DMs (2026)

An Instagram MCP server is a tool that lets AI assistants like Claude, ChatGPT, or Cursor control your Instagram Business account through Anthropic's Model Context Protocol (MCP). With it, the AI can read your DMs, respond to leads, set up comment-to-DM triggers, manage your CRM, and orchestrate Instagram workflows — all through natural language.

This is one of the newest niches in the AI tooling space. Instagram has historically been the hardest social platform to automate because Meta requires going through their official Business API, which is gated and complex. As of 2026, almost no commercial Instagram MCP servers exist — except SetSmart, which exposes ~180 MCP tools covering the entire Instagram DM lifecycle.

This guide covers what Instagram MCP is, what's available, and how to use it.

What Is an Instagram MCP Server?

The Model Context Protocol (MCP) is an open standard from Anthropic (released late 2024) that lets AI models connect to external tools through a unified interface. Before MCP, every AI tool needed custom integrations for every service. With MCP, a single server can expose any service's functionality to any MCP-compatible AI client.

An Instagram MCP server is an MCP server that exposes Instagram functionality:

• Read incoming DMs and conversation history
• Send replies, voice notes, story replies
• Manage comment-to-DM triggers and keyword automations
• Pull lead lists, qualification status, booking data
• Update Instagram CRM (tags, notes, lead status)
• Run analytics on conversation rates, qualification rates, booked calls

Once connected, you can ask Claude things like:

• "Pull my Instagram DMs from the last 24 hours that haven't been replied to. Draft responses I can review."
• "Add the keyword 'PLAN' as a comment-to-DM trigger that sends my free meal plan PDF."
• "Show me which Instagram leads commented on my Reel from yesterday but didn't qualify. What objections did they raise?"
• "Disable Instagram AI for the next 3 hours, I want to handle a VIP prospect personally."

The AI executes through MCP tool calls. You don't write code or click through dashboards.

Why an Instagram MCP Server Matters

Instagram DMs are the highest-volume sales channel for coaches, consultants, and agencies in 2026. Our study of 828K AI conversations found that Instagram generates more qualified leads in absolute numbers than WhatsApp (40,490 vs 35,669) thanks to sheer volume — but managing them at scale is brutal.

The problem: Instagram conversations are asynchronous, fragmented, and high-volume. A growing creator might have 200 active conversations across DMs, comments, and Story replies. Manually managing them is impossible. Even with automation tools like ManyChat, you're locked into rigid flows and can't easily query, segment, or orchestrate at the meta-level.

The solution: An Instagram MCP server gives you AI-level orchestration on top of your DM automation. You're no longer clicking through a dashboard — you're describing outcomes and the AI executes.

What's Available Today

Unlike WhatsApp MCP — which has 5+ open-source options on GitHub — the Instagram MCP space is nearly empty. The reason: Instagram's official Messaging API (via Meta) is gated behind business verification and developer review. You can't just spin up an Instagram MCP from a personal account.

A few experimental projects exist:

• gumloop/instagram-mcp — A workflow node, not a full MCP server. Limited to specific automation steps within Gumloop's platform.
• Various GitHub experiments — Personal projects that scrape Instagram via headless browsers. Not sustainable (Instagram blocks them quickly), violate TOS, and put accounts at ban risk.

The only commercial Instagram MCP server in production is SetSmart.

SetSmart Instagram MCP: ~180 Tools for Instagram DMs

SetSmart is an AI DM setter that already automates Instagram lead qualification through GPT-level AI conversations. The MCP layer sits on top — exposing every feature of the platform as MCP tools that Claude, ChatGPT, or Cursor can call.

What's exposed via MCP

Instagram account management:

• instagram.get_status — connection status, account name, AI on/off
• instagram.get_connect_instructions — onboarding URL for adding accounts
• instagram.disconnect — disconnect an Instagram account
• instagram.get_profile — pull Instagram Business profile data

DM triggers and keywords:

• dm_trigger.list — list all comment-to-DM keyword triggers
• dm_trigger.set_keywords — replace the keyword list
• dm_trigger.add_keywords / dm_trigger.remove_keywords — manage individual keywords
• dm_trigger.set_strict_mode — exact match vs fuzzy match
• dm_trigger.test — simulate a comment to test the trigger

Comment auto-replies:

• comment_reply.get_auto_reply — current public reply template
• comment_reply.set_auto_reply — update the public reply

Conversations:

• conversation.list / conversation.search — search by status, channel, date, text
• conversation.get — full message history
• conversation.bulk_tag — tag multiple conversations
• conversation.export — export conversation data

Contacts:

• contact.find / contact.list — search by name, Instagram handle, tags
• contact.import_from_attachment — bulk import from CSV/XLSX
• contact.add_tag / contact.remove_tag — segmentation

AI control:

• ai.pause_channel — pause AI on Instagram (e.g. "pause for 2 hours")
• ai.resume_channel — resume AI handling
• ai.update_blacklist — exclude specific contacts

Follow-ups:

• follow_up.list — pending follow-ups
• follow_up.send_now — trigger a manual follow-up
• follow_up.cancel — cancel a scheduled follow-up

Analytics:

• analytics.dashboard — overall Instagram stats
• analytics.channel_breakdown — Instagram-specific metrics
• analytics.daily_stats — qualification + booking by day
• analytics.get_persona — AI's understanding of the user's typical lead persona

Bookings (when connected):

• calendly.list_event_types / book_calendly_appointment
• ghl.list_calendars / book_ghl_appointment
• iclosed.list_events / book_iclosed_appointment

Cross-channel (multi-channel agents):

• whatsapp.* tools (full WhatsApp MCP coverage)
• messenger.* tools

Example: Instagram lead triage with Claude

You're a fitness coach with 20K Instagram followers. You post a Reel and get 80 comments. You open Claude Desktop:

"What's my Instagram performance from yesterday's Reel?"

Claude calls analytics.daily_stats filtered to yesterday. Returns: 80 comments, 47 triggered DMs (via your "PLAN" keyword), 22 qualified leads, 6 booked calls.

"Of the 22 qualified leads, which ones haven't been booked yet?"

Claude calls conversation.search with qualified=true, booked=false, channel=instagram, date_range=24h. Returns 16 conversations.

"Show me their last messages and what they asked about."

Claude calls conversation.get for each. Summarizes: "8 asked about pricing, 5 asked about timeline, 3 mentioned they're undecided about commitment level."

"Send a personal voice note follow-up to the 8 pricing questions. I'll record it once and SetSmart should send to all of them."

Claude prompts you to record a 30-second voice note. You record it. Claude calls whatsapp.send_voice_note (or Instagram equivalent) for each contact. Reports: "8 voice notes sent successfully."

That's a workflow that would take 30 minutes manually. With the MCP, it's 90 seconds.

How to Connect SetSmart MCP to Claude

1. Sign up for SetSmart at https://setsmart.io (7-day free trial)
2. Connect Instagram through Meta's official OAuth flow (60 seconds)
3. Get your MCP URL in the SetSmart dashboard
4. Add to Claude Desktop:
   • Open Settings → MCP servers
   • Add the SetSmart MCP server URL
   • Authenticate with your SetSmart credentials
5. Use it — Claude now has all 180 Instagram, WhatsApp, and Messenger tools

You can also connect to:

• ChatGPT via custom GPT actions
• Cursor via ~/.cursor/mcp.json config
• Any MCP-compatible client — the protocol is open

Use Cases for Instagram MCP

1. Daily Instagram lead triage

Set up a recurring prompt: "Every morning, show me yesterday's Instagram activity, identify the top 5 qualified leads I should personally reach out to, and draft a custom message for each."

2. Comment-to-DM campaign management

"I'm launching a new lead magnet next week. Set up the keyword 'GUIDE' as a comment-to-DM trigger that sends a PDF link, then asks 'What's your biggest challenge with X right now?' to start qualification."

Claude orchestrates dm_trigger.add_keywords + updates the AI script for the qualifying question + generates a confirmation report.

3. Cross-channel lead routing

"For all leads who first messaged me on Instagram and gave me their phone number, move the conversation to WhatsApp where my booking rate is 3x higher. Send them a friendly transition message."

Claude pulls Instagram conversations with phone data, sends a WhatsApp opener, and tracks the migration.

4. Performance analysis

"Compare my Instagram qualification rate this month vs last month. Tell me what changed and which content drove the most qualified leads."

Claude pulls analytics.daily_stats, joins with the AI's persona data, and highlights patterns.

5. Manual takeover

"Pause Instagram AI for the next 3 hours. I want to handle the inbound from my livestream personally. Resume after at 6pm."

Claude calls ai.pause_channel, schedules a resume_at task, and confirms.

6. Audit and cleanup

"Find all Instagram contacts I haven't messaged in 90+ days and aren't tagged as 'client'. Tag them 'cold' so I can run a re-engagement campaign."

Claude queries contact.list with date filters and contact.add_tag in bulk.

What the Data Says About Instagram DM Engagement

From our analysis of 373,042 Instagram DM conversations on the SetSmart platform:

• 17.77% of engaged Instagram leads qualify (vs 34% on WhatsApp)
• 6.28% of engaged Instagram leads book a call
• Median messages to qualify on Instagram: 23 (~11 back-and-forth exchanges)
• A single follow-up nearly triples Instagram qualification (+182%)
• 53% of conversations die before message 3 — the engagement cliff is real
• Instagram generates the most qualified leads in absolute numbers (40,490) thanks to massive volume

For the full benchmark study, see our 828K conversation analysis.

Instagram MCP vs Traditional Instagram Automation

Traditional Instagram automation tools are great if you want a visual builder. The MCP approach is better if you prefer AI-driven orchestration — describing what you want and letting the AI execute.

For a deeper comparison, see our Instagram DM automation guide or ManyChat alternatives breakdown.

Security & Best Practices

Use the official Meta API. Any Instagram MCP that scrapes Instagram via headless browsers will get accounts banned. SetSmart connects through Meta's Instagram Messaging Platform — the official API.

Approve destructive actions. SetSmart's agent prompt requires explicit confirm=true for destructive operations (delete contacts, cancel subscription, disconnect account). Always review what Claude proposes before approving.

Watch for scope creep. With 180 tools, the AI has broad access. If you're using a shared Claude workspace, be careful what conversations pass through.

Token security. Your Instagram access token is the equivalent of your password. SetSmart stores it encrypted server-side; the MCP only ever returns masked previews. Never copy raw tokens into chat history.

FAQ

What is an Instagram MCP server?

An Instagram MCP server is a tool that connects Instagram to AI assistants like Claude or ChatGPT through Anthropic's Model Context Protocol. The AI can read DMs, send replies, manage comment-to-DM triggers, run analytics, and orchestrate Instagram workflows through natural language. It removes the need to write code or click through automation dashboards.

Is there a free Instagram MCP server?

A few open-source experiments exist on GitHub, but none are production-ready or safe for business accounts. They scrape Instagram via headless browsers (which violates Instagram's TOS and risks bans). The only commercial, business-safe Instagram MCP server in 2026 is SetSmart, which uses the official Meta Instagram Messaging API and starts at $99/month.

Can I use Instagram MCP with ChatGPT?

Yes. MCP is a protocol standard, not Claude-specific. SetSmart's Instagram MCP works with Claude Desktop, ChatGPT (via custom GPT actions), Cursor, and any MCP-compatible client.

Will an Instagram MCP get my account banned?

Not if it uses the official Meta API. Open-source MCPs that scrape Instagram via Puppeteer or browser automation violate Instagram's Terms of Service and will get accounts restricted. SetSmart connects via Meta's official Instagram Business API — no ban risk, full Meta partner status.

What can an Instagram MCP automate that ManyChat can't?

ManyChat is a visual flow builder — you click to create automations. An Instagram MCP server lets the AI execute on natural language. You can ask Claude to "find all leads who mentioned X but didn't qualify, segment them by objection type, and draft personalized re-engagement messages" — that's not possible in ManyChat. Plus, the SetSmart MCP includes AI conversation handling that outperforms flow-based bots by 27x in lead qualification.

Does Instagram MCP work for multiple Instagram accounts?

Yes. SetSmart supports multi-account setups — useful for agencies managing several client accounts. Each account is connected to your SetSmart workspace, and the MCP can target a specific account via parameters.

How does Instagram MCP differ from WhatsApp MCP?

The MCP protocol is the same. The tools are different — Instagram MCP exposes comment-to-DM triggers, Story replies, and Reels-specific automation. WhatsApp MCP exposes broadcasts, templates, and click-to-WhatsApp ad referrals. SetSmart's MCP includes both, plus Messenger, in a single server.