Descriptions:
Matthew Berman shares an exhaustive look at how he runs OpenClaw—the open-source personal AI assistant—as a full-time, 24/7 productivity system on a dedicated MacBook Air. The setup is accessible remotely via Tailscale SSH and TeamViewer, integrates models from Anthropic, Google Gemini, xAI Grok, and OpenAI, and uses a hybrid SQL-plus-vector database to support both structured queries and natural language search across all stored data.
One of the most technically detailed sections covers his Twitter/X data pipeline, which chains four API tiers by cost and capability: FX Twitter (free, single-tweet lookup only), TwitterAPI.io (search and thread context at $0.15 per 1,000 tweets), the official X API v2 ($0.005 per tweet, pay-per-use, full access), and xAI’s Grok as a final fallback. The system automatically selects the cheapest path for each query type.
Berman also walks through YouTube analytics tracking with nightly competitor scans, a business meta-analysis system inspired by Coinbase CEO Brian Armstrong—where a council of eight parallel AI agents analyzes all available business data and delivers ranked recommendations every night—plus a personal CRM pulling from Gmail and calendar integrations. Each workflow section includes the actual natural language prompts used to build the capability, making the entire system reproducible by viewers without writing code directly.
📺 Source: Matthew Berman · Published February 11, 2026
🏷️ Format: Workflow Case Study
