This week was about quiet refinement — the kind of changes you feel every day without thinking about them. We shipped a new model, made the chat thread nicer to live in, gave you a faster way to share feedback, and added a new plan for larger teams. Here's the tour.
GLM 5.2 model support
GLM 5.2 is now built into Zero. It's supported natively — it shows up in the model list with the Built-in badge, so you can pick it and start working on your VM0 credits with no API key or setup required.
Prefer to bring your own key? That works too: connect a direct Z.AI API key and your routes default to GLM 5.2 (with GLM 5.1 still available whenever you want it), or use your own key through OpenRouter. Either way, pricing and provider routing are already wired up, and your existing default and selected models are untouched — nothing changes until you choose to switch.
Your chat drafts now stick around
Start typing a message, get pulled into another thread, come back later — your unsent text is right where you left it. Zero now saves the draft in your composer per agent and per thread, so a half-written prompt survives navigating away, switching agents, or closing the tab.
There's nothing to turn on. Drafts save quietly as you type and clear themselves the moment you send. It's a small thing that removes a small, daily annoyance.
Double-click to rename a chat
Renaming a thread used to take a few clicks. Now you just double-click its title in the sidebar, type the new name, and you're done. Quick to reach, quick to fix, no menu hunting.
Inline feedback, right where you're reading
Select any text inside one of Zero's replies and a small toolbar appears: Copy and Provide feedback.
The real win is precision. Instead of typing a paragraph to explain which part you mean, you point straight at the exact lines and leave your note right there — pinpoint feedback, no context-setting required. You can even mark up several passages at once and send them together, so a single pass covers every spot that needs a change. (Copy, of course, grabs exactly the text you highlighted.)
It's now on for everyone, in every chat thread.
A new Enterprise plan
The pricing page now has a third option. Alongside Pro and Team, there's an Enterprise plan built for larger organizations: advanced security, custom credit volume and concurrent runs, unlimited agents, bring-your-own LLM keys, dedicated support with an SLA, and volume and annual billing. Pricing is custom — the card's Contact sales button starts the conversation.
If your team has outgrown self-serve, this is the path that flexes to fit. Pro and Team are exactly as they were.
From the Blog
GLM-5.2 Joins Zero for Long-Context Agent Work — a closer look at where GLM-5.2 shines in Zero: long-context coding, project-scale refactors, debugging, and tool-heavy agent workflows.
BI Without a Data Lake — run privacy-safe founder BI directly on a masked Neon database, no data lake or warehouse required. Ask questions in plain language and get answers.
Join us in San Francisco
Meet & Build: Run Your Company With AI
Wednesday, June 24, 6:00–8:00 PM PDT at Frontier Tower (10th Floor Annex), 995 Market St, San Francisco. An in-person evening for solo founders and operators: swap notes on the operational grind, watch a live Zero demo, and stick around for an optional hands-on session to sketch your first AI workflows. Hosted by the VM0 team.
That's everything we shipped this week. Most of it is the kind of polish you only notice when it's missing — try it out, and let us know what you think with the new inline feedback toolbar.
