Getting Started
Introduction
What the WhatsApp Automation Platform does, the full scope of its capabilities, and what you can build with it.
What is this platform?
The WhatsApp Automation Platform gives you an AI agent that lives on your WhatsApp number. It doesn't just answer questions — it can browse the internet, write and run code, generate images, automate websites, collect payments, schedule messages, learn new skills on its own, and connect to virtually any external service.
When someone sends a message to your WhatsApp number, the AI agent reads it, understands the intent, and takes action — whether that's looking something up online, running a Python script to analyze data, filling out a form on a website, generating a custom image, triggering an M-Pesa payment, or scheduling a follow-up for next week.
The only limit is your imagination. If you can describe it, the agent can probably do it.
What can the agent do?
Browse the web
The agent has full, unrestricted internet access. It can search for information, read web pages, and bring the results into your conversation. Ask it to look up a competitor's pricing, find a news article, research a topic, or check any website — it goes and gets the information in real time.
Automate websites
The agent has its own persistent browser — a real Chrome instance that maintains login sessions, cookies, and history. It can navigate websites, fill out forms, click buttons, log into accounts, extract data from dashboards, and automate multi-step workflows on any website. If you log into a service through the agent's browser, it stays logged in for future tasks.
It can also learn and replay browser automations. Once it successfully completes a browser task, it saves the workflow as a reusable skill that it can run again later — faster and more reliably.
Write and run code
The agent can spin up persistent code sandboxes to write and execute Python, TypeScript, or JavaScript. It can process data, run calculations, generate files, build scripts, and return the results. Need a CSV analyzed, a chart generated, or a script that transforms data? The agent writes the code, runs it, and sends you the output — all within the WhatsApp conversation.
Files created in the sandbox can be exported and sent back to you as WhatsApp attachments.
Generate and edit images
With AI image generation built in, the agent can create images from text descriptions, edit existing images, and upscale low-resolution photos. Ask it to create a product mockup, design a poster, edit a photo, or generate any visual content you need.
Collect payments
The M-Pesa integration lets the agent send payment prompts directly to customers' phones during a conversation. It handles the full payment flow — requesting payment, confirming receipt, issuing refunds, and processing payouts.
Schedule messages and follow-ups
Send messages at a specific time, set up recurring broadcasts to contact groups, or have the agent schedule its own follow-up reminders. The agent can create AI-generated broadcasts where each recipient gets a personalized message.
Learn and remember
The agent builds its own knowledge over time. It can save skills — procedures, workflows, and how-tos — that it learns during conversations and reuse them later. It can save notes for context it needs to remember. And it searches your Knowledge Base to ground its answers in your specific documents and data.
Delegate to sub-agents
For complex tasks, the agent can spawn sub-agents that work in parallel. Need research done on three different topics at once? The agent delegates to three sub-agents, each with the full toolkit, and combines the results. This makes it dramatically faster at multi-step tasks.
Connect to anything
Through the plugin system, you can connect the agent to any external service — CRMs, project management tools, email platforms, databases, or any API. Plugins extend the agent with custom tools that it can call during conversations. If a plugin doesn't exist for your service, you can build one.
Who is this for?
- Business owners who want an AI employee on WhatsApp that handles customer conversations, takes payments, and runs tasks 24/7
- Teams that need automated workflows — from answering support questions to filling out forms on external platforms to generating reports
- Developers who want to extend the agent with custom plugins and integrations
- Anyone who wants a personal AI assistant on WhatsApp that can actually do things, not just talk
Use cases
The platform is designed to be general-purpose. Here are some examples, but these barely scratch the surface:
- Customer support — Answer questions using your uploaded documents, escalate to humans when needed
- Sales automation — Qualify leads, send quotes, collect payments, and follow up automatically
- Research assistant — Search the web, read articles, summarize findings, and compile reports
- Data processing — Upload spreadsheets, run analysis in code sandboxes, generate charts and reports
- Social media management — Log into platforms through the browser, post content, extract analytics
- Appointment scheduling — Check availability, book appointments, send reminders
- E-commerce — Handle orders, process M-Pesa payments, track deliveries
- Content creation — Generate images, write copy, create documents
- Website monitoring — Periodically check websites for changes, prices, or availability
- Internal tools — Automate repetitive tasks that involve multiple websites, databases, or APIs
Key concepts
Instance — A connected WhatsApp number. Each instance has its own AI agent, settings, and capabilities. You can run multiple instances.
Agent — The AI assistant on an instance. You control its behavior through a system prompt, choose which AI model powers it, and decide which tools it can access.
Knowledge Base — Documents, text, and web pages you upload for the agent to search through. The agent uses this to give accurate, grounded answers specific to your business.
Toolkits — Groups of capabilities you can enable or disable. Some are built in (web browsing, code execution, skills), others can be installed (payments, scheduling, image generation).
Plugins — Custom extensions built by developers that connect the agent to external services and APIs.
Escalation — When the agent hands a conversation to a human team member for manual handling.
How to use these docs
Start with the Quick Start guide to get your first agent running in minutes. Then explore the sections in order — each page builds on what came before.
The Getting Started section covers the basics. Your AI Agent shows you how to configure the agent's behavior and knowledge. Conversations covers managing chats and contacts. Automation covers scheduling, payments, and integrations. Developers is for building custom plugins.