Skip to content
SocialMate
Sign in Download free
Presence · memory · agent tools

Every WhatsApp AI agent can send a message. Yours can read the room.

Most WhatsApp AI agents can only send a message. Yours can react, show it is typing, mark messages read, reply to a specific message — and start every conversation already knowing the whole history. Self-hosted, your own number.

Download SocialMate free Read the docs
In short

SocialMate gives your AI agent the things people actually do on WhatsApp, each as a callable tool: react to a message with an emoji, show a typing indicator before answering, mark a chat as read, quote and reply to one specific message, send a poll, a location or a contact card. Because they are tools rather than workflow steps, the model chooses when to use them — it reacts because the message deserved a reaction, not because step 4 always fires. And it starts every conversation already knowing the history: one Get AI Context call returns the whole thread as ready-made memory, including the years that happened before you built the agent. Self-hosted on your own machine and your own number, with no WhatsApp Business API.

  1. Mark Read — it acknowledges before it answers, so nobody is left watching an unread badge.
  2. React — the model decides this message deserved a reaction. Costs no message quota.
  3. Send Typing — an answer that arrives after a visible pause reads as considered, not automated.
  4. Get AI Context — it quotes a message from months before the agent existed. One call, no vector store.

Make an agent that behaves like a person

  1. Start from the message, not the schedule. A SocialMate Trigger on message.received (or an MCP client polling for new messages) hands your model the incoming message the moment it lands.
  2. Give it the history first. Attach Get AI Context as a tool. One call returns the thread as a role-mapped transcript, so the agent answers knowing what was already said — including conversations from long before the agent existed. Requires Pro.
  3. Let it acknowledge. Attach React and Mark Read. A thumbs-up on the message that only needed acknowledging is a reply that costs nobody a notification.
  4. Let it think out loud. Attach Send Typing (composing, or recording for a voice note). An answer that arrives after a visible pause reads as considered; one that arrives in 200ms reads as a machine.
  5. Let it answer the right message. Use replyTo on Send Text so the answer quotes the message it belongs to — the difference between a conversation and a stream of statements.
  6. Keep it inside the pacing engine. Nothing here bypasses anti-ban. Sends still pass rate limits, jitter, warming and risk checks; presence signals ride a separate lane so acknowledging someone never spends your daily message budget.

How SocialMate compares

Capability SocialMate This is us WAHA Evolution API Twilio (Cloud API)
Self-hosted, your own number Yes Yes Yes No
No WhatsApp Business API / approval Yes Yes Yes No
No per-message fees Yes Yes Yes No
Native n8n node (usableAsTool) Yes Partial Partial Yes
Native MCP server Yes Yes No No
Conversation-memory tool for agents Yes No No No
Reactions, typing and read receipts as agent tools Yes Partial Partial No
General-purpose AI assistant allowed Yes Yes Yes No — barred since 15 Jan 2026
Anti-ban engine Yes No No n/a
Smart queue + scheduling Yes Partial Partial Partial

SocialMate automates the WhatsApp app on your own machine. It is not the official WhatsApp Business API and is not affiliated with WhatsApp or Meta. Provider capabilities change — verify current details with each provider.

Frequently asked questions

Can an AI agent react to a WhatsApp message with an emoji?
Yes. React is a first-class tool on both surfaces — the n8n node's Message → React operation and the whatsapp_react_message MCP tool, both backed by the same REST endpoint. Because it is exposed as an agent tool rather than a fixed workflow step, the model decides which messages deserve a reaction. It works on the Free tier.
Can it show a typing indicator before it replies?
Yes — Send Typing sets presence to composing (or recording, for a voice note), and paused clears it. Pair it with SocialMate's human-like send delays and an answer arrives the way a person's would: a pause, a typing indicator, then the message. Also on the Free tier.
Does the agent remember previous conversations?
Yes, and this is the part that differs most from the usual setup. A typical n8n WhatsApp agent stores memory in a Simple Memory node or a vector store you build — it starts empty and only knows what happened after you switched it on. SocialMate syncs and stores your WhatsApp archive locally, so Get AI Context returns the real thread, including years that predate the agent. Reading that archive is a Pro feature.
Do reactions and typing count against my message limits?
No. Presence signals — reactions, typing, read receipts — are logged on a separate lane from messages, so they do not consume your per-minute, per-hour or daily send budget. Actual messages always do, and always pass the anti-ban engine first.
Why can't I do this on the official WhatsApp Cloud API?
Partly capability and partly terms. The Cloud API works on a verified WhatsApp Business number rather than your own personal one, and it has no access to the conversation history that existed before you onboarded. Separately, Meta's Business Solution terms effective 15 January 2026 do not allow providers whose primary product is a general-purpose AI assistant to use the platform — structured support, booking and notification bots are still permitted. SocialMate is not the Business Solution; it drives your own linked device.
Does SocialMate write the replies?
No, and that is deliberate. SocialMate is the hands and the memory; your own model — OpenAI, Claude, anything you can attach in n8n or point at the MCP server — is the brain. AI text generated by SocialMate itself is on the roadmap, not shipped.
Is behaving more like a human a way around WhatsApp's rules?
No — and it should not be sold as one. These are ordinary WhatsApp features that people use every day, exposed to your agent so conversations feel like conversations. Automating any account can still breach WhatsApp's terms no matter how natural it looks; message people who expect to hear from you, warm new numbers slowly, and treat the anti-ban engine as a floor rather than a guarantee.
Also see Ready to wire it up? See the WhatsApp AI agent tool for n8n →

Give your agent a presence, not just an outbox

Download SocialMate free, link your own number, and attach the tools to whichever model you already use. No card, no Meta approval, no per-message fees.

Download free Compare Free vs Pro
Free forever · no card Download free