What we shipped, week by week. Subscribe to get it in your inbox.
Agents can now choose an app's visibility (public, unlisted, private) on the first publish. When publishing changes a live app's visibility, the message tells the agent exactly what changed (e.g. "The app is now unlisted (it was public)"). Agents cannot publish private apps.
When app AI calls run out, viewers now see a clear quota notice (shown once per app view). The upgrade CTA appears only for personal workspaces, and visitors of public apps can now use AI against their own workspace. Per-visitor rate limits (20/min, 200/hour per app) prevent a single visitor from draining a workspace's quota.
A new Data tab section lists recent agent data changes with one-click undo. Every workspace member can see and revert agent changes from a sidebar entry, and undo respects collection write rules and preserves original row authors.
A new Connect your agent button on the dashboard header opens the setup steps and Feldova's MCP address in a dialog, so connecting an agent is always one click away. The dashboard onboarding card is now an agent setup guide.
New users can now sign up and verify their email while connecting an agent in a single continuous flow — through email/password, Google, and the verify-email page. Signed-in links in the sign-up/sign-in flow preserve pending agent connections and invites.
Auth paths now reject open-redirect attacks using backslashes or control characters, and agent tokens whose scopes exceed the user's approved consent are properly rejected on reconnect.
OAuth client registrations that never completed a connection within 7 days are now automatically removed.
App emails are now capped per workspace per day (100 on Free, 1,000 on Team). Scheduled emails that hit the cap aren't retried, and apps receive a clear DAILY_CAP_REACHED error code so they can respond gracefully.
Legal and documentation now reflect agent-built apps, new plan limits, no creator payouts, and clearer disclosures about data flows to your agent's provider. The agent guide has been rewritten for agents only with a dedicated usage-limits section and restored file-layout rules.
After an agent publishes, the studio now updates its publish controls instantly instead of still showing "Publish" or reopening the first-publish dialog.
When an agent retries with an expired client registration, users now see a friendly "This agent connection is out of date" page instead of a bare error. Expired sign-in signatures now automatically restart the agent connection with a fresh signature instead of showing a misleading "Invalid email or password" error.
Runtime AI calls now count against a simple monthly quota per plan instead of per-call credits. Credit packs, auto-refill, referrals, signup grants, creator payouts and billing history are gone. Billing shows AI usage directly, and plans are priced by AI calls, email, and team features.
Agents can now request one-time upload URLs, attach files to ingest content with metadata, and discover which of your apps accept ingest. Uploads are retryable and the ingest dialog leads with your connected agent, keeping tokens out of copied text for script-only use.
The New app dialog now simply asks for a name — no AI prompt required. Your connected agent takes it from there.
The studio now shows a live feed of versions your agent saves, with the preview reloading automatically as new versions come in. The old in-app AI chat has been replaced by an agent panel focused on what your connected agent is doing.
Grant your agent data:read and data:write scopes to let it list collections, read rows, write rows, run server actions, and add ingest content on your behalf. Every data change is journaled as an undoable batch with conflict checks, so you can safely revert anything your agent did.
Agents can now validate an unchanged draft (validate is a dry run and no longer refuses no-op edits). The lint no longer incorrectly flags all-caps words like "FAQ" inside JSX text, which previously blocked agent saves.
Our homepage now leads with "Your AI agent builds it. Feldova runs it." with a new hero, repositioned how-it-works page centered on connecting your agent, publishing, and working with app data, and updated plan pricing. Public header and footer link straight to Connect your agent.
Dashboard onboarding steps for "agent saved a version" and "owns a published app" now check across every workspace you own, not just your personal workspace. A failed check now reports as not done instead of crashing the whole dashboard.
The mobile app now ships with a full-bleed brand icon on a violet gradient instead of a mark floating on a white square, so it stands out on your home screen. The Android 12+ splash logo is also sized to sit cleanly inside the circular mask instead of getting clipped.
The workspace switcher on mobile now slides up from the bottom while the backdrop fades in — and reverses smoothly on close — instead of the whole dim sliding up the screen.
Flipping "Allow anyone with the link" in the share dialog now shows a toast confirming the new access state. Previously the toggle felt like a no-op because the link URL itself never changes — only who can open it.
On the mobile Apps tab, long-press any app tile to open a native menu with a Remove from workspace action. A normal tap still launches the app as usual. Removals are reversible — you can always reinstall from Browse — and non-admins see a clear message if they don't have permission.
The mobile launcher now uses an even 4-column grid that fills the screen width edge-to-edge, with tile and icon sizes that scale cleanly across phones and tablets. App icons also render with the same gradient used on the web, so tiles look vibrant instead of flat.
Installing an app from Browse now updates the Apps tab right away instead of showing a stale grid until you reopen it. The installed apps list is shared across screens so every view stays in sync.
Tab and app screen titles are now centered on Android to match iOS, the Apps header's + button no longer sits flush against the screen edge, the Browse back button is properly labeled "Apps", and opening an app uses the native back button with swipe-to-go-back.
Sticky footers in dialogs now reach the rounded bottom edge cleanly, instead of appearing to float above a strip of scrolling content.
Mobile builds now pin the correct Node and pnpm versions so production builds no longer fail from a mismatched lockfile, and Android production builds are submitted to the closed-testing track that gates release.
Pasting long unbroken text like console errors or blob URLs into a studio chat bubble no longer pushes the bubble past the edge of the panel. Long tokens now wrap while preserving stack-trace indentation.
We retired the legacy frontend assembly path in favor of the modern esbuild-based bundler everywhere — seed apps, examples, and live apps. This fixes a class of "app already declared" and blank-app failures (including some content ACL and comments regressions), and ensures apps with only a bare App function still render correctly.
Apps created before we started tracking platform versions are once again flagged as outdated, so you can update them to the latest platform standards with one click.
Buttons, inputs, selects, checkboxes, radios, and switches now grow to a comfortable 44px touch target on mobile and shrink back to a compact size on desktop. Dialog footers stick to the bottom of the screen on mobile with full-width actions, making forms and confirmations far easier to use on phones.
When your app writes to shared storage, we now automatically snapshot the previous value before every update or delete. If something important gets overwritten or removed, admins can browse workspaces, apps, and rows in a new data recovery page and restore any version from the last 30 days (up to the 10 most recent per key). Old revisions are cleaned up nightly.
Attaching an image to a chat used to silently drop or hard-fail on generation models that can't natively read images. We now describe each attached image once with a vision model and pass the description through, so image-aware conversations keep working across every model.
Failed forks used to return an opaque 500 with no server-side trace. Permission and not-found failures now surface as proper 403/404 responses, and unexpected errors are logged so they can actually be diagnosed.
The Android tab bar mixed tinted line icons with a full-color 🔔 emoji for notifications. All tabs now render as matching single-color line icons in both active and inactive states. iOS is unchanged.
Shared collection rows now carry updatedAt and updatedBy end to end, so apps can badge rows that have changed since a viewer last saw them. Anonymous, link-grant, and scheduled writes intentionally leave updatedBy unset so per-viewer "unseen" logic can't be fooled by system-attributed writes.
In the app overflow menu, clicking Edit on an app you weren't a member of would silently fork it with no indication a copy was being made. The fork action is now spelled out as "Fork & edit," while "Edit" only ever opens the studio on apps you can edit directly.
The "‹ Apps" back control sat flush against the app title and read as one run-on label. It now has its own spacing and a slightly larger tap target.
The mobile app now ships with the real stacked-squares brand mark across the iOS launcher, Android adaptive icon (including the themed monochrome variant), splash screen, and Play Store listing — replacing the earlier placeholder chevron.
Release builds of the mobile app were failing to bundle JavaScript due to a missing Babel preset. Production Android builds now complete successfully.
Setting an unprefixed width like max-w-2xl or max-w-5xl on a dialog was silently ignored on desktop, leaving it stuck at 384px. Dialogs now widen as requested, both in generated apps and across the platform (screenshots, chat gallery, feature catalog, and more).
The native app grid and browse lists now show each app's actual icon — the same gradient tile with lucide glyph you see on the web — instead of a plain colored box with the first letter of the name.
Fork failures now return meaningful status codes (403 for permission issues, 404 for missing apps) instead of a generic 500, and unexpected errors are logged so problems can actually be diagnosed.
Fixed a missing dependency that caused production Android builds to fail during JavaScript bundling. Release builds now complete reliably.
Added breathing room between the back button and the app title so the back control reads as its own button — and gets a slightly larger tap target.
Setting a wider max-width on a dialog (like max-w-2xl or max-w-5xl) previously had no visible effect on desktop — dialogs stayed capped at 384px. Custom widths now apply correctly, both in generated apps and across platform dialogs like screenshots, chat gallery, and the feature catalog.
Replaced the mismatched color emoji notification icon on Android with a clean line icon, so all tabs now render as consistent single-color icons that respect the active/inactive tint. iOS is unchanged.
When you don't have direct edit access to an app, the overflow menu now clearly shows "Fork & edit" instead of a generic "Edit" item, so it's obvious when clicking will create a copy. Members of the app's workspace still see a plain "Edit" that opens the studio directly.
Publishing an app — or swapping in new code under an already-public app — now runs the name, description, frontend, and server code through an automated content classifier. Clearly disallowed content is blocked with a specific, appealable reason; borderline cases are flagged for a human. Suspended creators can no longer push new code to existing public apps to work around a suspension.