Buzzably
Changelog

Release Notes

See what's new and what we've improved.

103 releases

v0.7.1

September 2, 2026
🐛Fixed1
  • Behind-the-scenes improvements to how articles are published and how their details, such as reading time, are recorded. No action is needed on your part.

v0.7.0

September 1, 2026
✨Added1
  • Articles now show an estimated reading time, so you know what you're getting into before you start.
🔄Improved2
  • General improvements across the platform's interface, accessibility, and performance.
  • Studio Home now shows a live status widget for your AI agents, so you can see what's running, what needs attention, and what's finished without leaving the page.
🐛Fixed1
  • Ongoing security hardening across the platform. No action is needed on your part.

v0.6.0

August 14, 2026
✨Added2
  • Links to Buzzably shared on social platforms now show our on-brand preview card.
  • You can now share any insight directly to your favorite platforms with one tap, right from the reader view.
🐛Fixed1
  • Ongoing security hardening across the platform. No action is needed on your part.

v0.5.0

August 1, 2026
✨Added1
  • Pro plans can now also run the option previously available only on Free. Economy drafts on a smaller AI model and uses fewer tokens per run, so it is a lighter, quicker pass when a draft does not need the full model. Standard remains the default, and you can switch per draft.
🔄Improved2
  • Small interface and translation improvements across the app.
  • Public pages and posts now load faster, and your light or dark theme applies immediately instead of briefly flashing the wrong one.
🐛Fixed2
  • Pausing a scheduled AI agent now reliably stops it from running.
  • Times across Content Management and the editor — including when a post is scheduled to publish — were shown in one fixed regional time zone for everyone, regardless of where you are. They now follow your own time zone, so the time you set is the time you meant.
🔒Security1
  • Ongoing security hardening across the platform. No action is needed on your part.

v0.4.0

July 14, 2026
✨Added2
  • Pro subscribers' Draft Writer and Persona Stylist now run on Claude Sonnet 5, a substantially more capable model. Free-tier writing moves to GPT-5.6 Luna. You should notice better structure, tone-matching and long-form coherence — with no change to how you use the assistant. Persona word targets are now also set per plan (Free 800 words, Pro 2,000, Enterprise unlimited), so the AI writes to a length your plan supports instead of silently over-running.
  • The platform now supports a wider range of AI models across providers, each tuned to be used the way it works best. Models are selected for you based on your plan.
🔄Improved1
  • The Studio sidebar has been reordered to follow the order you actually work in.
🐛Fixed18
  • If you navigated away, refreshed, or hit a connection blip while a draft was generating, the finished draft could be stranded and never shown to you. The assistant now picks up where it left off and surfaces the completed draft when you return.
  • A gating error could block Free-plan users from the AI Draft Assistant they were entitled to. Corrected.
  • The scheduling screen showed times in UTC with no indication, making it easy to schedule for the wrong hour. Times now appear in your own timezone.
  • Auto-generated titles, teasers and tags came back in English even when your draft was written in Traditional Chinese. They now follow the language of your content.
  • Trial information attached during invite-based sign-up could be incomplete. Corrected.
  • On pages that front-load navigation and boilerplate, the length cap could cut away the article and keep the menus. The cap now preserves the actual article text.
  • Invite links now carry a time-limited token that is exchanged when you click it, making acceptance both more reliable and more secure.
  • Content scheduled directly from the editor could be missed by the scheduler. Fixed.
  • Some invite-based Pro trials were not applied to the account. Fixed — and you are now notified when your trial begins.
  • Long AI drafts could be truncated part-way through and fail to save. Requests are now shaped for each model individually, so the full draft is generated and returned intact.
  • The step-by-step progress indicator could freeze part-way through even though the draft was still generating normally. Progress now tracks the real state of the run.
  • Links in welcome emails could point to the wrong address in some cases. Corrected.
  • Saving your settings could wrongly report your own username as "already taken". Fixed.
  • Importing a URL as a source could fail silently or return partial content. It now retries, degrades gracefully, and — when a page genuinely cannot be read — tells you why instead of failing quietly.
  • Scheduled content could publish late. Publishing is now event-driven and fires at the time you set.
  • Occasionally a draft would fail at the final step with an internal formatting error — most often on longer pieces and on Chinese-language content. The AI's output format is now guaranteed by the provider, so this class of failure is gone.
  • Following an invite with an account that already exists showed a confusing error. You now get a clear message and a direct route to sign in.
  • The waiting state could go blank or appear stuck. It now always shows meaningful progress and recovers cleanly from interruptions.
🔒Security1
  • Documents you upload to train a persona are now scanned and structurally validated before they are processed, with clearer errors when a file cannot be accepted.