
What's new in Zero — Week of May 12
This week in Zero: Personal Models hits GA, 11 new connectors arrive, and built-in generation gets faster. See every product update from the week of May 12, 2026.
Insights, tutorials, and updates on AI teammates and the future of human-AI collaboration.

This week in Zero: Personal Models hits GA, 11 new connectors arrive, and built-in generation gets faster. See every product update from the week of May 12, 2026.

We ran 22 user interviews on workflow automation and found four barriers no one talks about. Here is why automating your workflow is still so hard, and what helps.

AI agent security starts with one rule: your agent should never hold your credentials. See how a broker pattern keeps your tokens safe, using a 38-year-old idea.

This week in Zero: drop almost any file into chat, smarter image uploads, and a sharper gallery view. See every product update from the week of May 5, 2026.

How I built an AI workflow with Zero to automate our blog production end to end, from draft to illustration to publish. A practical look at agent-run content ops.

This week in Zero: a Telegram connector, the new Usage hub, audio in chat, and a cheaper DeepSeek V4 Pro. See every product update from the week of April 28, 2026.

A frontend engineer's take on design-as-code: how a designer and one engineer rebuilt our entire platform in 12 days using an AI-era design-engineering workflow.

This week in Zero: six new features, 19 fresh use cases, and $100 in free credits to try them. See every product update shipped the week of April 21, 2026.

The copilot era is plateauing. See what changes when AI operates autonomously as a colleague, the capabilities a true AI teammate needs, and why it matters now.

Zero now runs on Claude Opus 4.7 across vm0, at the same price with upgraded reasoning and capability. See what the model upgrade means for your agent workflows.

We raised $14M. Here is what Zero actually is, what our team of six built in a single month, and where the product goes next. A candid build update from VM0.

Why your agent's system prompt acts like a bureaucracy, and what context, not control, really means for building AI agents that adapt instead of just following rules.