WhatsApp AI Agents With AWS
Budget / Salary$3,000–5,000
TypeFreelance project
LocationRemote
Posted2 hours ago
I need three Claude-powered AI agents running inside a single WhatsApp integration and writing to my existing AWS database.
Scope
• One shared WhatsApp entry point. The Claude model should automatically detect whether the incoming message is:
– a general customer question,
– data we need to capture or update in the database, or
– a guided workflow request such as booking, ordering, or support.
• Each detected intent is handled by its own isolated agent, but all agents must reuse common code where practical and log every interaction to the same AWS tables.
• Claude’s code base is already licensed to me; please wire your functions into it rather than replacing it.
Technical requirements
• WhatsApp Business Cloud API for messaging.
• AWS Lambda (Python preferred) or an equally serverless solution that connects to my current RDS/PostgreSQL instance.
• Clean, well-documented code checked into my Git repository.
Acceptance criteria
1. End-to-end demo: sending three sample messages triggers the correct agents and records entries in the database.
2. Accurate field mapping between WhatsApp payload and existing table schema.
3. Readme explaining environment variables, deployment steps, and how to add future agents.
Scope
• One shared WhatsApp entry point. The Claude model should automatically detect whether the incoming message is:
– a general customer question,
– data we need to capture or update in the database, or
– a guided workflow request such as booking, ordering, or support.
• Each detected intent is handled by its own isolated agent, but all agents must reuse common code where practical and log every interaction to the same AWS tables.
• Claude’s code base is already licensed to me; please wire your functions into it rather than replacing it.
Technical requirements
• WhatsApp Business Cloud API for messaging.
• AWS Lambda (Python preferred) or an equally serverless solution that connects to my current RDS/PostgreSQL instance.
• Clean, well-documented code checked into my Git repository.
Acceptance criteria
1. End-to-end demo: sending three sample messages triggers the correct agents and records entries in the database.
2. Accurate field mapping between WhatsApp payload and existing table schema.
3. Readme explaining environment variables, deployment steps, and how to add future agents.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.