Changelog
What we shipped this month: launch edition
Run Pilots is live: connected accounts, scheduled flights, MCP catalog and approvals.
Read post →Blog
Notes from the toweragents, browser, MCP
Product updates, engineering notes and practical guides for scheduled automation.
Security
Prompt injection and the agentic web
What we do when a web page tries to give your agent instructions.
Read post →Engineering
Playwright vs Selenium vs browser clouds
The honest trade-offs behind our browser layer choice.
Read post →Customers
Case study: an agency posting 200 times a week
How a five-person agency replaced three ops tools with scheduled browser flights.
Read post →Engineering
Cutting agent costs with model routing and step budgets
Simple levers that took our average run from 48 credits to 11.
Read post →Product
Human-in-the-loop: when agents should ask permission
Autonomy is a dial, not a switch. Where we place approval gates and why.
Read post →Guides
Your first autonomous flight in five minutes
A concrete walkthrough: from signup to a scheduled post that runs while you sleep.
Read post →MCP
MCP for operators: connect any tool to your agent
The Model Context Protocol is not just for developers. Here is what it means for the tools you already use.
Read post →Engineering
How we keep X sessions alive securely
Persistent logins without keeping passwords: envelope encryption, isolation and health checks.
Read post →Product
Why browser agents beat APIs for the long tail of the web
APIs cover the popular 10% of the work. The other 90% lives behind a login form — and that is exactly where browser agents shine.
Read post →Subscribe via RSS.