Native MCP, not a REST wrapper.
Waaru exposes documented workspace operations as MCP tools with explicit inputs, scopes, and results. The same product rules still apply whether a person or an approved agent starts the action.
Native MCP server
Connect AI tools to your WhatsApp workspace through MCP.
Waaru ships a native Model Context Protocol server that exposes the full WhatsApp workspace surface as MCP tools. Your AI agent - Claude, GPT, Gemini, or anything that speaks MCP - operates Waaru directly. Send messages, manage contacts, run broadcasts, fetch analytics, escalate to humans, edit resolution memory, all through standardized MCP tool calls.
Platform free. AI features are paid separately.
Expose tools with explicit boundaries.
Inspect the tool name, required workspace scope, and result before an external agent can use a WhatsApp operation.
ExamplePermissioned MCP tools
Workspace-scoped request
- Tool
- contacts.find
- Scope
- contacts:read
- Result
- Contact found
Tools the agent can call.
- send_message - single contact or list
- send_template - Meta-approved templates with variables
- create_contact, update_contact, tag_contact
- create_broadcast, schedule_broadcast, broadcast_report
- list_conversations, get_conversation, update_conversation_status
- search_knowledge - Company Brain search
- list_resolutions, add_resolution, update_resolution
- escalate_to_human - with note and priority
- create_payment_link - Razorpay-backed
- schedule_visit - Cal.com / Google Calendar
Auth - workspace-scoped, revocable, segmentable.
Per-workspace MCP keys. Issue multiple keys per workspace with different tool allowlists - read-only analytics, full service surface, broadcasts only. Revoke from the dashboard.
Example agent workflows.
Your existing AI agent becomes a WhatsApp operator. No custom API integration code. Customer-service AIs reply to messages directly. Sales agents qualify leads through your existing agent infrastructure. Analytics agents pull conversation stats on demand.
FAQ
Frequently asked questions.
What is MCP (Model Context Protocol)?
An open protocol from Anthropic standardizing how AI agents discover and call tools across services. AI agent → MCP server → executable tool, with typed arguments and standard responses.
Which AI agents can use Waaru's MCP?
Any MCP-compatible client: Claude Desktop, Cursor, Zed, and custom agents built on the MCP SDK (Python, TypeScript, Rust). The agent connects to Waaru's MCP endpoint with a workspace API key.
Is MCP secure?
MCP security depends on the server implementation and deployment. Waaru uses workspace-scoped, revocable keys, tool allowlists, server-side workspace isolation, and call logs. Teams should still grant only the tools an agent needs and require human approval for high-impact actions.
What plan includes MCP?
Public signup is open. The platform is free. AI features are paid separately. Meta messaging charges remain separate.
Compare
Evaluating platforms? Every claim cited, every price dated.
Ready to ship a WhatsApp flow that doesn't break?
Public signup is open. The platform is free. AI features are paid separately.