One accent, six phase colours and the same motion quality on every tier — the pipeline now reads as one product instead of twenty. Faster runs when Apple Intelligence assets aren't ready: the app now stops retrying a model it knows can't answer, instead of repeating the same failed call for every agent. Prompt copies now expire from your clipboard after five minutes and never sync to other devices. Restore Purchases no longer tells a paying Lifetime member they have nothing when the App Store is briefly unreachable. Also: per-agent generation timeouts so one stalled agent can't hang a run, and clearer disclosure that the SEO Blog Generator is a separate purchase.

Meta Prompt Lab 2.4.9 • Fixed a completion crash reported by beta testers • Pipeline animation now stays synchronized with live steps • Improved navigation spacing across the screen

Meta Prompt Lab This update introduces paywall and ratings engine optimisations: • Contextual Access: Added intelligent prompts for daily limit thresholds, Pro features, and day-three returns • Entry Indicators: Clear display of remaining runs today so you can plan your builds • Lapsed Restore: Streamlined screen to easily re-subscribe and restore Pro access • Ratings Engine: Improved feedback requests to only ask for ratings after multiple successful builds All private processing still runs 100% on your device.

Meta Prompt Lab This update introduces paywall and ratings engine optimisations: • Contextual Access: Added intelligent prompts for daily limit thresholds, Pro features, and day-three returns • Entry Indicators: Clear display of remaining runs today so you can plan your builds • Lapsed Restore: Streamlined screen to easily re-subscribe and restore Pro access • Ratings Engine: Improved feedback requests to only ask for ratings after multiple successful builds All private processing still runs 100% on your device.

Meta Prompt Lab Blog generation is more reliable on device. The pipeline now keeps sections within practical word-count bounds, expands thin sections more consistently, and shows clearer live progress while writing. This update also improves prompt comparison reliability and includes release-readiness fixes for iOS 26.5.

Meta Prompt Lab Polish based directly on your beta feedback: • Completed runs now show a clear "done" badge — no more mistaking the finished state for loading • Live activity feed refined — accent-lit panel that tracks the working agent in real time • Tab bar now tucks away as you scroll, giving your content the full screen • Fixed a missing icon on the Privacy & Trust panel in Settings • Cleaner enhanced-prompt output — removed awkward double-negative phrasing in rules Everything still runs 100% on your iPhone. Nothing leaves your device.

Meta Prompt Lab A big stability and polish release: • Fixed a crash that could interrupt a run right before your enhanced prompt finished • History: opening a saved prompt now has a proper Back button • All-new live pipeline screen — glowing orb visualization with a real-time agent activity feed • New: SEO Blog Generator — draft SEO-optimised articles fully on-device • Smoother animations, faster and more dependable 20-agent runs Everything still runs 100% on your iPhone. Nothing leaves your device.

Faster, smoother, and friendlier. A quicker path to your first expert prompt, a cleaner upgrade experience, and a more polished first run. Turn any rough idea into a structured, ready-to-use prompt for ChatGPT, Claude and Gemini — 100% on-device, 100% private.

Faster, smoother, and friendlier. A quicker path to your first expert prompt, a cleaner upgrade experience, and a more polished first run. Turn any rough idea into a structured, ready-to-use prompt for ChatGPT, Claude and Gemini — 100% on-device, 100% private.

Smoother, more efficient pipeline animation, plus battery and performance improvements under the hood. Always free to start — 100% on-device, private and offline.

What's new in 2.0.9: • Bigger, smoother pipeline animation — watch all 20 agents build your prompt on iPad and Mac. • A more premium Create screen: sharper glass, cleaner layout. • Smarter Refine chat that understands your prompt better. Always free to start. 100% on-device — private, offline, no account.

What's new in 2.0.8: • Bigger, smoother pipeline animation — watch all 20 agents build your prompt on iPad and Mac. • A more premium Create screen: sharper glass, cleaner layout. • Smarter Refine chat that understands your prompt better. Always free to start. 100% on-device — private, offline, no account.

• Refine without limits — keep chatting with on-device Apple Intelligence to reshape any prompt as much as you want. • A smoother, more responsive Refine chat that streams replies as they arrive, like a real conversation. • Visual polish on the Create screen.

Refined for iOS 26. • A deeper, more tactile Liquid Glass design across every screen, with smoother motion and richer detail. • Calmer, faster pipeline visuals — and a cleaner experience when a run can't complete. • Tapping a saved prompt in History now reliably opens its full details. • A range of polish, accessibility, and reliability improvements.

New: a value card appears right after your pipeline finishes to show you the biggest win from your session. New users can run a sample prompt in one tap during onboarding — no faffing about, straight in. We've also fixed a bug that could prevent subscription details loading on slow connections. It now loads reliably every time.

This update fixes a crash some people could hit when starting a pipeline run, so your prompts now build smoothly and reliably every single time. We also refined the live animation: a single calm, glowing orb, softer energy arcs that grow into each step, and a deeper, more focused background. Steadier, cleaner, and easier to follow from start to finish.

Refined the running pipeline animation for a cleaner, more focused look. A single bright orb now anchors the screen, with a full ring of nodes that lights up in clockwise order as each step completes. Electric arcs reach out and grow into every node, so you can watch your prompt take shape in real time. Smoother, brighter, and easier to follow from start to finish.

Reimagined the live pipeline: a cinematic plasma orb with lightning that strikes out to each AI agent as it runs. Smoother, faster, and more premium throughout. Plus a refreshed Refine screen and polish across the app.

Premium running screen overhaul. The Living Orb is rebuilt with a swirling nebula interior, dense starfield, a bright pinpoint core, and an energy tendril that connects the active agent to the orb in real time. Constellation arms now grow as each agent completes its work. Smoother pipeline log animations and refined glass header. Reduce Motion and Low Power Mode fully respected.

v2.0.0 is a complete visual overhaul. New Living Orb pipeline hero, refreshed Liquid Glass surfaces, smoother per-agent handoffs, and tightened motion across every screen. The on-device privacy promise is unchanged: 20 agents, your prompt, no cloud.

This release sharpens Liquid Glass across the app and quietly extends battery life. Highlights - Liquid Glass refinements across Create, Refine, History, and Vault - Battery-aware animations: when iOS Low Power Mode is on, animations dial back automatically - Accessibility polish: improved VoiceOver labels and tighter focus paths - Snappier history loading and clearer locked-row affordances Thanks for using MetaPromptLab. As always, your prompts never leave your device.

v1.9.91 Polish release. Same 20-agent on-device pipeline, smoother every step. Premium pipeline motion Pipeline DAG view now renders curved Bézier edges, particle comet trails, hot-core ignition flashes, and staggered halo ripples as each of the 20 agents completes. Free and Pro tiers see different motion intensities, calibrated for clarity at small screen sizes. Refined app icon App icon refreshed with a higher-contrast, more vivid version that reads cleanly at every size from Spotlight thumb to large Home Screen tile. Tier-aware agent counter Pipeline now displays 8 of 8 agents for free users and 20 of 20 agents for Pro and Lifetime subscribers, with a live odometer-style count that ticks as each agent completes. Smarter preset prompts Tapping a preset prompt now replaces the input field instead of appending. Rapid double-taps are filtered to prevent accidental concatenation of two different presets. Stability and reliability Animation runtime hardened: deterministic gradient stop ordering, zero-width frame guards, smoother frame pacing on iPhone 12-class devices. Privacy unchanged All 20 agents still run entirely on your device. No prompt text, no personal data, no analytics payload leaves your iPhone, iPad, or Mac.

v1.9.9 Polish release. Same Apple Intelligence pipeline, smoother every step. Premium pipeline motion Pipeline DAG view now renders curved Bézier edges, particle comet trails, hot-core ignition flashes, and staggered halo ripples as each of the 20 agents completes. Free and Pro tiers see different motion intensities, calibrated for clarity at small screen sizes. Refined app icon App icon refreshed with a higher-contrast, more vivid version that reads cleanly at every size from Spotlight thumb to large Home Screen tile. Stability and reliability Animation runtime hardened: AgentTimeline gradient stops now ordered deterministically, zero-width frame dimensions guarded, motion tokens tuned for smoother frame pacing on iPhone 12-class devices. Privacy unchanged All 20 agents still run entirely on your device. No prompt text, no personal data, no analytics payload leaves your iPhone, iPad, or Mac. This is an architectural constraint, not a policy.

v1.9.8 Follow-up release polishing what v1.9.7 introduced — based directly on TestFlight + App Store reviewer feedback. Live Activities, refined Dynamic Island progress now updates more smoothly during long pipeline runs. The agent-by-agent counter is easier to read at a glance, and the final-score handoff back to the app is faster. On-device classifier, v2 The pre-pipeline prompt-quality classifier has been retrained on a wider distribution of UK English inputs. Expect noticeably better early-flagging on dissertation outlines, coursework briefs, CV bullet points, and A-level / GCSE revision prompts. Paywall + premium path Micro-copy on the subscribe screen has been tightened for UK consumer-law clarity. The 3-day free trial messaging, auto-renewal disclosure, and cancel-anytime line are now unambiguous. Restore-purchase is one tap. Prompt Vault polish Faster scrolling on large vaults, better tag autocomplete, and a fix for a rare SwiftData write race when saving immediately after a pipeline run completes. Quality gates + scoring Minor tuning on G10 (open-ended directive) and the signal-to-noise dimension — Pro tier users will see slightly stricter validation on long-form prompts, which is exactly what reviewer feedback asked for. Privacy note Nothing has changed. All 20 agents still run entirely on-device. No prompt text, no personal data, no analytics leave your phone or tablet. That is an architectural guarantee, not a policy. What is coming next iCloud sync for the Prompt Vault, a Pro referral mechanic, and three onboarding flows tuned for writing, coding, and research. No committed dates — we ship when it is right. Thank you to everyone who filed feedback on Live Activities timing and the classifier. v1.9.8 is shaped directly by those reports. Tap to install — free. Pro anytime.

v1.9.7 Cleaner paywall. Smarter scoring. Live updates. Paywall layout The subscribe screen has been redesigned so the Subscribe button is the primary action, with auto-renewal disclosure sitting below the fold. This matches Apple and UK consumer law guidelines and makes the upgrade decision cleaner for everyone. Live Activities and Dynamic Island MetaPromptLab now pushes real-time pipeline progress to the Dynamic Island and Lock Screen during a run. Watch individual agents complete as your prompt is being engineered — no need to stay in the app. On-device prompt-quality classifier A new lightweight classifier pre-screens your input before the 20-agent pipeline starts. It flags ambiguous or under-specified prompts early, so the pipeline has better material to work with and your quality scores improve from the first run. Apple Search Ads improvements SKAdNetwork v3 conversion-value patches are now live. These are invisible to you as a user but mean the app can accurately measure which keywords lead to Pro subscribers — which in turn lets us invest the Apple Search Ads budget in channels that actually bring in engaged users. Premium plan path Several small friction points on the upgrade and settings flow have been resolved. The route from free tier to Pro is now shorter, and subscription state reflects immediately after purchase. Stability and reliability Post-ship code review from v1.9.6 has been addressed: minor edge cases in JSON parsing, QualityScorer tie-breaking, and SwiftData history writes are all resolved. Privacy note Nothing has changed here. All 20 agents run entirely on your device via Apple Intelligence (FoundationModels). No prompt text, no personal data, no analytics payload leaves your iPhone or iPad. That is a hard architectural constraint, not a policy. What is coming next The next train will bring iCloud sync for your prompt vault, a referral mechanic for Pro members, and three new onboarding flows tuned for different use cases (writing, coding, research). No committed dates — we ship when it is right. Thank you to everyone on TestFlight who filed feedback on the paywall layout and Live Activities timing. v1.9.7 is directly shaped by those reports. Tap to install — free. Pro anytime.