Build AI agents on WhatsApp in less than a minute!

Description and use

Wappy lets you quickly build a personal AI agent that runs on WhatsApp and stores data locally, cutting setup to about a minute.

Problem

Creating a personal AI agent that runs on a user’s own machine, uses their data, and communicates via WhatsApp was difficult because of WhatsApp plumbing, memory persistence, and tool integration.

Outcome

Wappy reduces the setup time for a WhatsApp‑based AI agent to about a minute for scaffolding and a few additional steps for deployment, enabling users to run agents locally with their own data.

Workflow

• Install Node.js 22+
• Run npm create @wappy_ai/agent to scaffold an agent
• Answer the generator prompts
• Set up a Meta Cloud API app and webhook as guided
• Choose a model (Anthropic, OpenAI, Gemini, or Ollama)
• Configure memory (SQLite/LibSQL or Cognee)
• Ingest knowledge base for grounded retrieval
• Deploy the agent on your machine
• Interact via WhatsApp

Evidence

Wappy is a free, open‑source (MIT) TypeScript SDK for building your own AI agent on WhatsApp. It handles the plumbing so you can focus on the part that makes your agent yours.

Source

The generator asks a few questions and scaffolds your agent in about a minute.

Source

Your agent, its memory and your data live on your machine. WhatsApp messages still pass through Meta, and if you choose a hosted model, it receives your prompts.

Source