What’s new in Playcode

Give a teammate Admin on a single project and they can now run it end to end.

Publish, snapshot, restore. A project admin can publish and unpublish the project, create and restore checkpoints, retry provisioning, and manage its published subdomain - without needing to be a workspace admin.

Destructive actions stay protected. Deleting or moving a project, and attaching custom domains, still require the project's creator or a workspace admin. So you can hand someone the keys to operate a project without risking that it gets erased or relocated.

Workspaces now have their own billing. A team workspace can subscribe to Pro or Business, buy credits, and manage its payment method - independently of anyone's personal plan.

Every workspace has its own credit balance. Work done in a team project uses the team's credits, not yours - and a teammate on a free personal plan gets the full power of the workspace's plan while building there, including the best AI models. Your personal workspace keeps its own plan and credits, exactly as before.

Everyone sees the team's credit meter. Members can always check how many credits the workspace has left and when they refresh. Payment details stay private: only the workspace owner and admins can see the card, change the plan, buy credits, or configure auto-recharge - members who run low see who to ask instead of a dead end.

Upgrading is contextual. Open pricing from inside a team workspace and the subscription goes to the team; open it from your personal workspace and it's yours. Each workspace gets its own invoices and billing portal.

Existing subscriptions carried over automatically - if you were subscribed, your personal workspace has the same plan, credits, card, and invoice history as before, with no action needed.

Playcode now has workspaces - shared spaces where your whole team builds together.

Create or switch workspaces from the main dashboard menu and invite teammates by email - they don't need a Playcode account yet; the invitation walks them through signing up. On phones, projects, folders, and the workspace menu share one sidebar, so navigation never stacks drawers. Everyone in the workspace sees its projects live: when a teammate creates or edits a project, it appears for you in real time.

Sharing got personal, too. The Share popup in the editor now works the way you'd expect from tools like Notion: type an email, pick a role (Viewer or Editor), and that one person gets access to that one project - whether or not they're in your workspace. Projects shared with you show up in a new "Shared with me" section in the sidebar.

Viewer access is consistently read-only across code, chat, publishing, and Cloud controls. Editors can change project content without receiving owner-only sharing or lifecycle controls, and any direct collaborator can leave a shared project from the Share popup.

Each workspace has its own settings area: manage People (roles, pending invitations), rename or leave the workspace, inspect Usage per member, and open Billing as its own section. Owners and admins can filter usage by period, inspect individual AI requests, and understand input, output, and cache credit use. Billing controls stay role-aware: owners and admins manage the workspace plan while members get a clear read-only explanation. Paid features and Upgrade prompts follow the active workspace, so teammates receive the plan their team is actually paying for.

Your personal projects are untouched - everything you've built lives in your personal workspace, exactly as before. And when a project should belong to the team, Move to workspace in the project menu brings it - chats, history, and all.

Playcode's AI model picker just got the newest models from OpenAI and xAI, plus a cleanup so the list stays easy to choose from.

OpenAI GPT-5.6 arrives as three tiers you can pick per task:

• GPT-5.6 Sol - the flagship. Highest quality of the family, for the hardest builds.
• GPT-5.6 Terra - the balanced everyday pick. GPT-5.5-level quality at lower credit use, so it's the new default GPT for most work.
• GPT-5.6 Luna - fast and light. Great for simple pages and quick edits where you'd rather iterate than wait.

Grok 4.5 from xAI joins too - an "Opus-class" model that's fast and unusually token-efficient, so long AI builds tend to cost less.

Each model's picker badge reflects its expected credit use, so you can trade quality for cost with your eyes open. Start on Terra, reach for Sol when a build is genuinely hard, and drop to Luna or Grok 4.5 for quick, cheap passes.

We also retired older models from the picker (earlier GPT-5.x, Grok 4.1, and a few others) so the choices you see are the ones worth using. Nothing you've already built is affected.

The projects dashboard now lets you work on many projects at once - a real help once you have more than a handful.

On desktop, drag a box across empty space to rubber-band a group of projects (it auto-scrolls when you reach the edge), or hold Shift to extend a range. On a phone, tap the checkbox on each card. Once you've selected some, a bar appears with the actions that fit: Move to folder, Archive, or - for already-archived projects - Restore and Delete permanently.

We tidied the dashboard around it, too. Each card now shows whether it's a Static or Cloud project, the filter row carries live counts for your Cloud usage and your Archive, and the views - All, Live, Drafts, Cloud, Archive - moved into one clear switcher at the top while the sidebar stays focused on your folders.

Ask Playcode's AI to check your site's speed and it now runs a real Google PageSpeed Insights (Lighthouse) audit on your live page - the same signals Google uses to rank sites.

You get your Performance, Accessibility, Best Practices, and SEO scores, your Core Web Vitals (Largest Contentful Paint, Cumulative Layout Shift, and more), and a ranked list of concrete fixes with estimated time savings. The AI can measure, apply improvements, and measure again to prove your site actually got faster.

Publish your site first, then just ask - for example, "How fast is my site, and how do I speed it up?"

It's a step toward Playcode helping your site succeed, not just get built: fast pages that are ready for Google, all in the same chat.

We rebuilt the storage system behind Playcode Cloud - the layer that keeps your full-stack app's data and files safe while it runs.

The payoff is more predictable performance and stronger reliability, especially for apps that read and write a lot of data. Most of it is invisible by design: your app just runs more smoothly, with fewer surprises under load.

Nothing to change on your side. Existing Cloud apps get it automatically.

Your Cloud terminal now keeps running as long as its tab is open.

Before, switching to another tab and back would drop the session and start a fresh shell - losing whatever was running and your command history. Now you can move between the terminal, preview, and AI Chat freely: the terminal stays connected in the background, and switching back brings you right back to the same session, scrollback and history intact. Dragging the terminal to another panel keeps it too.

Closing the terminal tab still ends the session. And if you leave the whole app for a while, idle terminals disconnect on their own to keep things tidy - reconnecting the moment you come back.

Claude Fable 5 is now selectable in Playcode's AI model picker.

Fable 5 is Anthropic's highest-capability generally available Claude model for demanding reasoning, coding, and long-horizon agent work. It has much higher credit use than Sonnet and Opus models, so the picker shows it as a premium option for the hardest tasks.

Claude Sonnet 5 credit estimates now account for its launch pricing and newer tokenizer together.

The model picker badge and actual credit use are aligned, so Sonnet 5 should no longer overstate expected credits during the introductory pricing window.

Pro mode now starts new chats at Extra High effort when you have not chosen a different default.

This gives the default Quality model more room to reason through multi-step build and debugging work. You can still choose Low, Medium, High, Extra High, or Max from the Intelligence setting when you want to trade speed and credit use for depth.

Claude Sonnet 5 is now the default Quality model in Playcode.

New Pro chats use Sonnet 5 for the Quality path by default. Existing users who were following the previous default move with the Quality setting automatically, so they do not need to change anything in Preferences.

Opus 4.6 remains available in the model picker for work where you explicitly want Opus.

Claude Sonnet 5 is now selectable in Playcode's AI model picker.

It is a new Sonnet option for builders who want a stronger balanced model without jumping all the way to Opus. Sonnet 5 uses Anthropic's newer tokenizer, and the model picker badge accounts for expected credit use.

Opus 4.6 remains the default Quality model for now while we watch Sonnet 5 on real website-building work.

Pro Quality mode now runs directly on Claude Opus 4.6 with extended thinking turned on. It's noticeably more stable and consistent on longer, multi-step work, and the model is shown by name in the AI Model picker and your usage history.

Nothing to change on your side - Quality just gets the upgrade. The Intelligence dial still tunes how hard it thinks per task, and Economy mode stays the fast, lightweight lane.

When you ask the AI to recreate a design from a screenshot, check how a page looks, or work with an image, it now sees the actual picture - not just a link to it.

Before, the AI could take a screenshot but sometimes couldn't really look at it, so it would capture the same page over and over without making progress. That's fixed. Now it looks once and gets to work.

What the AI can see now:

• Screenshots it takes of a live page - so it can clone a design you point it at, or check that your published site looks right.
• Images it generates or edits - it sees its own result, so it can judge it and refine instead of working blind.
• Any image you give it a link to - paste an image URL and ask the AI to describe it, match its style, or build from it.

It also accepts more image formats automatically (including AVIF) and resizes large images for you, so you don't have to convert anything first.

Pro mode now lets you choose how much the AI thinks before it answers. Open the mode picker in the chat and set an Intelligence level for the task in front of you:

• Low / Medium - quick answers for simple edits, with less waiting
• High (the default) - the balanced setting that fits most work
• Extra High / Max - let the model reason longer on hard, multi-step problems

Set it per chat from the mode picker, or choose a default for every new chat in Preferences > AI Model > Default Intelligence. Economy mode stays the fast, simple lane, so there's no dial to think about there.

Also in this release

• No more blank Pro replies - on rare occasions a Pro answer could come back empty; those now retry automatically instead of showing nothing.

</content> </invoke>

Picking Premium mode in the AI chat sometimes did nothing in a freshly opened project - the switch quietly snapped back to Economy. That's fixed: your mode choice now sticks in every project, even before you've sent your first message.

Improvements

• Premium mode is selectable everywhere - the choice is saved the moment you make it, in any project
• Set your default mode - choose whether new chats start in Economy or Premium from Preferences > AI Model > Default Mode, and it applies across all your projects

Pointing your own domain at an app is the step where most people get stuck. So we rebuilt it into a guided setup wizard for Playcode Cloud apps that walks you through it, step by step.

Open Custom domains in the publish menu, type the domain you own, and the wizard does the hard part:

• One record, not a pile of them. It detects your DNS provider, shows the single record to copy, and the exact name your provider expects. No TXT ownership record, no IP addresses, no certificate to install.
• Ownership and SSL are automatic. Once your record resolves, Playcode confirms the domain is yours and issues the HTTPS certificate for you.
• Close the window and walk away. Setup can take a few minutes for DNS to update, so we email you the moment your domain is live instead of making you wait and refresh.
• Root or www, handled. Want yourdomain.com without the www? The wizard shows the right record when your provider supports it, and points you to www when it doesn't.

Each step has a plain-language Need help? panel answering the questions first-timers actually ask, so you never need to know what a CNAME is to finish.

See the full walkthrough in the custom domain help guide.

Two billing fixes for existing subscribers. Opening the Billing Portal silently failed on mobile (it worked on desktop) - the link is now opened in a way phones don't block. And changing your plan (for example Pro → Business) no longer dead-ends on "redirecting to secure payment" - subscribers are now sent to the billing portal's plan switcher.

Fixes

• Billing Portal opens on mobile - previously nothing happened on phones; the portal now opens reliably
• Plan upgrades work for current subscribers - switching plans routes to the billing portal instead of failing

When Playcode's AI takes on a multi-step job (setting things up, running your tests, building, or publishing), it now tells you what it's doing in plain language as it happens, instead of showing raw technical output. You'll see what it's checking and why, like "Are there already settings files here?", and how a longer step is going, like "Tests about halfway, passing so far", so you're never left watching a blank, frozen screen.

On longer jobs the AI now works more like a teammate at the keyboard: it starts the step, keeps an eye on it, and gives you a quick read every few seconds, flagging early if something looks stuck instead of going quiet until the very end. Clearer, simpler, and written for founders, not engineers.

Sites served on your own custom domain now load over HTTP/2. Browsers can fetch many files at once over a single connection instead of requesting them one after another, which is the same modern protocol your visitors already get on Playcode subdomains.

The effect is most noticeable on mobile, where pages have many scripts and styles to load. In our mobile testing on a custom-domain site:

• Mobile Lighthouse performance: 73 → 85
• Largest Contentful Paint: 4.7s → 3.5s

No setup is needed. Every site published to a custom domain on Playcode Cloud gets this automatically.

Cloud publishing now observes deploy logs in short, bounded checks instead of waiting on one long-running server step. This helps the publish dialog keep up with newer publish and rollback activity, especially when a build step is quiet for a while.

If a publish is cancelled or a deploy step takes longer than expected, Playcode can now notice follow-up publish activity faster instead of appearing stuck with no output.

Cloud project previews now come back much faster after they have gone idle.

For projects with a warm snapshot available, Preview can wake in about 1-2 seconds instead of waiting for a full cold start that can take about a minute. That is up to around 60x faster when you return to a recent Cloud project.

This is most visible when you are iterating on a full-stack app: open the project, wait for Preview, make a change, leave it idle, then come back later. Playcode can now resume the working project state instead of rebuilding the whole dev environment every time.

If a project needs a clean start, for example after an infrastructure update or an expired snapshot, Playcode falls back to the slower cold start path so the project still opens correctly.

GLM 5.2 is now selectable in Playcode's AI model picker. It replaces GLM 5.1 in the visible picker and gives you a newer long-context reasoning option for code-heavy tasks.

Choose it from Menu → Preferences → AI Model when you want to compare GLM against your usual model on the same project.

GLM 5.1 is no longer shown in the picker. Existing history remains intact, and the newer GLM 5.2 option is the one to use going forward.

We fixed a caching issue that could keep the Preview showing old CSS after you changed a cloud project. Cloudflare was telling browsers to keep some Preview stylesheets for too long, so a normal reload could bring back stale styles instead of the latest version from your project.

New Preview requests now respect the project server's cache headers, so CSS changes refresh normally.

If your browser already cached the old response, clear it once: open the affected project or folder, open the browser developer console, go to the Network tab, turn on Disable cache, and reload the page. After that one reload, the stale Preview cache is cleared and future updates should work normally.

Playcode now gets to the usable editor faster. We split more of the heavy editor code into lazy-loaded pieces, load only the current language by default, and delay expensive startup work until the interface is already responsive.

Preview starts more gently. The browser no longer has to spin up the preview worker, bundler, and Service Worker immediately on page load. If you press Run, Preview still starts right away. If you just open a project and begin typing, the heavy work waits briefly so the page stays smooth.

Preview is also more reliable in embedded browsers. Playcode now checks for blocked browser features like Service Workers, cache storage, IndexedDB, cookies, and local storage. When a browser blocks something Preview truly needs, you get a clear message instead of a blank panel or a forever-loading iframe.

Sign-in and onboarding got a cleanup pass. The auth screens are cleaner, Google sign-in fallback behaves better when a browser blocks it, and new users get a smoother first-run flow before landing in the editor.

We also tuned the marketing homepage video so it waits for real user intent instead of competing with the first paint, which makes the first screen feel less stuck on slower devices.

You can now invite friends to Playcode and both of you get credits. Your friend receives 10 welcome credits the moment they sign up with your link, and you receive 40 credits when they subscribe.

Find your personal invite link in the project menu, right under your credits meter. You have 3 invite passes; every time an invited friend subscribes, one pass comes back, so your best invites never run out.

The invite dialog shows who joined and what you have earned, with friends' emails masked for privacy.

Usage History now shows where every AI credit goes. Click the info icon on any row for a full breakdown: input, output, and cache tokens, each with its own credit cost.

It answers the question we hear most: why did a short message still cost credits? The first message in a chat saves your whole prompt and tools to cache once (that is the big number), and every later message reuses it at a large discount. Now you can see exactly that.

The activity list also got a proper Load more button and a running count of how many records you are viewing, so you can browse your full history.

OpenAI's newest general-purpose model, GPT-5.5, is now available alongside 5.4 in Playcode's model picker. Pick it as your Quality model in Menu → Preferences → AI Model → Quality Model if you want to try it.

Heads up on cost. Per OpenAI's May 2026 rate card, GPT-5.5 is roughly 2× the per-token price of GPT-5.4 across input, cached input, and output. 5.5 shares the same tokenizer family as 5.4, so net effective per-request cost lands at about 2× what 5.4 charges for the same work. Real, not theoretical.

Not the default - Claude Sonnet 4.6 keeps that slot for now. We add new models as they ship so you can A/B them against your usual pick on real work; no model is unilaterally better than the others on every job, and the right answer depends on what you're building.

Honest take: no AI today gets everything right. Even Claude Opus, which still does UI design better than any model we've tested, takes its time and sometimes over-engineers a request. GPT-5.5 has its own trade-offs. Run the same prompt across two or three models when it matters; trust your own eyes on the output.

Earlier today we moved the default Quality model to Claude Opus 4.8. A few hours of real-world Pro usage made it clear that 4.8 burns credits roughly 2.6× faster than Claude Sonnet 4.6 for comparable output on typical website-building work.

Concretely: on the entry-level Pro plan, leaving 4.8 as the default would have torn through the monthly credit allowance in a handful of busy days. That's not the experience we want anyone to land on by default.

So we're walking it back: the default Quality model is now Claude Sonnet 4.6. Strong reasoning, sharp design, well-balanced per-request cost. Your Pro credit budget goes meaningfully further.

Opus 4.8 stays available for the work where the extra capability is worth the cost. Switch any time in Menu → Preferences → AI Model → Quality Model. Same place for Opus 4.6 if you want the previous default.

Honest about the back-and-forth: we set 4.8 as default thinking the quality jump would be worth the spend. The first hours of paid-user data said otherwise. Better to adjust quickly than to leave you on a setting that doesn't fit your plan.