Connect a local WhatsApp archive to Claude through a narrowly scoped MCP, with explicit participant privacy, retention, tool permissions, and read only defaults so message search does not quietly become unrestricted surveillance or sending.
Source episode · 29:47
How I AI: 3 AI Workflows to Save $140K & Automate Marketing with ElevenLabs' Luke Harrieswith Claire Vo
Luke runs an unofficial local WhatsApp bridge, authenticates it through the WhatsApp Web QR flow, stores synchronized messages in SQLite, exposes search and send capabilities through an MCP server, and connects Claude to summarize recent groups and search messages about a topic.
A local searchable message index with bounded MCP tools, clear sync status, source chat and timestamp citations, and sending disabled or tightly constrained until the owner intentionally enables it.
Anthropic AI assistant
Step by step
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
5 steps
Clone and inspect the bridge and MCP code, dependencies, permissions, storage paths, and exposed tools. Run it under a dedicated local user or directory, keep secrets and databases out of Git, and start with send disabled.
Authenticate through the WhatsApp Web QR flow on the owner's device. Confirm which chats and history sync, where SQLite and media are stored, how updates run, and how to revoke the linked session and delete local data.
Run the MCP server locally with read only tools first. Add authentication where supported, bind only to the intended interface, minimize logs, and restrict chat IDs, time range, result size, and attachment access.
Register the local server with Claude and inspect every discovered tool and schema. Disable or wrap broad send, delete, and attachment actions before granting message search access.
Query one allowed chat and short time range, requiring citations and explicit coverage. Expand only after confirming privacy and accuracy; draft outbound messages separately and send only through a deliberate constrained action.
Search only [allowed chat] from [start] to [end] for [topic]. Return a concise summary with source message IDs or timestamps, participants as permitted, disagreements, and missing coverage. Do not search other chats, expose unrelated content, or send any message.
Turn an idea into a PRD, user stories, and a plan.
After the steps
How to recover when the loop fails and where human judgment helps.
Recover
Keep building
Design a weekly meta-workflow that analyzes your interactions with your AI assistant to make it smarter. This system learns your writing style from edits, suggests new automations for repetitive tasks, and helps you filter valuable AI techniques from hype.
Create a daily AI workflow that summarizes yesterday's meetings, identifies urgent tasks, and proactively learns internal company terms it doesn't understand. This keeps your AI assistant's knowledge current and highly relevant to your work.
Build a recurring AI task that automatically gathers updates from your calendar, Slack, and notes to suggest weekly priorities and prepare you for upcoming meetings. Start every Monday with a clear, comprehensive plan without the manual effort.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.