Nalya (Persimmon · Immersive) — Ship the app. We grow it.

The problem

Building the app was the easy 20%.

Growth isn't one big task. It's a hundred small recurring ones. Skipping them sounds like this.

★☆☆☆☆user review

“worked great for a year, then the last update made it crash the second i hit save. lost my work twice now. left a review like 2 weeks ago, never heard anything back. done”

💬 A quick, specific reply actually pulls ratings back up. Most never get sent.

🧑‍💻indie dev

“not gonna lie, my release notes have just said ‘bug fixes and improvements’ for the last 14 updates. writing real ones in 25 languages every release is never gonna happen”

📝 A free conversion surface, thrown away on every single update.

😵‍💫2 a.m.

“woke up to like 180 one-star reviews. a crash shipped to 100% overnight while i was asleep and i had no idea. rating went 4.7 → 4.3 in a single day”

🧯 Crash rates should've halted that rollout before it reached everyone.

🔎solo founder

“spent a whole month wondering why installs were sliding. turned out a competitor outranked me on my own main keyword. found out way too late”

🔎 Keywords drift and rivals move constantly — search ranking is never “done”.

📊founder

“my analytics has been screaming ‘58% drop at the paywall’ since february. it's on my todo list. it's been on my todo list for four months”

📉 The data knows where they leave. Nobody has time to test the fix.

💸indie dev

“still charging $4.99 flat in every country. i know that's way off for brazil and india, but setting up regional prices is an afternoon i never seem to have”

💰 Regional pricing and offer tests are upside almost nobody captures.

🧩the growth stack

“my whole growth setup is 7 browser tabs and me copy-pasting between them. reviews in one, ranking in another, analytics somewhere else. i'm the integration”

🧩 Six dashboards that report. Not one of them acts.

How it works

A closed loop, not another dashboard.

Tools show you charts and stop. Nalya runs the whole cycle: it watches your store and product data, proposes the next best action, executes it through official store APIs, and measures whether it worked.

👁️Observestore stats, vitals, reviews, funnels

💡Suggestexperiments, ASO, replies, fixes

⚡Actvia official Play / ASC APIs

📐Measuredid it move the metric?

Your growth agent

Observe everything, continuously

Play install & rating exports, Android vitals (crash/ANR rates with 3-year history), every new review, plus your PostHog funnels — polled around the clock so nothing is missed.

Suggest with reasons, not noise

Every proposal arrives with the evidence behind it: “1-star reviews mentioning ‘crash on save’ tripled after 2.4.1 — halt the rollout and here's the stack trace.”

Act through the front door

All writes go through the official Google Play Developer API. No scraping, no gray areas, full audit log of every action taken.

Measure and learn

Each action is tracked against the metric it targeted. Replies that revise ratings, listing changes that lift conversion — the agent doubles down on what works for your app.

Everything included

Everything your launch needs. One connection.

Reviews, releases, ASO and revenue are live in beta — all driven from your editor over MCP, and all supervised first: you approve, it learns, then it earns autonomy. A/B tests, error triage and tester management are coming soon.

🚢

Publishing with auto release notes

No SDKv1

Push a build (or let us pull it from your GitHub Actions run) and Nalya handles the rest: localized release notes generated from your actual commits, track promotion, staged rollouts — and an automatic halt if crash or ANR rates spike against your baseline.

✓ Internal→ ✓ Closed→ ✓ Open beta→ ● Production · staged

Rollout at 25% · vitals gate passing — crash 0.09% vs 0.11% baseline · next step 50% in 24h

✨ Release notes generated from 14 commits

“New: pick a custom color for any pattern. Fixed the widget refresh on Android 16. Smoother wallpaper transitions on slow devices.”

en-USde-DEfr-FRes-419pt-BRja-JPko-KRhi-INru-RUtr-TRit-IT

💬

Review auto-replies

No SDKv1

Every review gets a specific, on-topic reply in the reviewer's language — grounded in your app's context and your past replies. 1-star crash reports also open a bug. Never templated, never spammy: store policy is built into the guardrails.

Marcus T. · Germany

★★★★★ · 2h ago

“App crashes every time I try to save a wallpaper since the last update. Pixel 9. Used to love it.”

🤖 Drafted reply · German

“Das tut uns leid, Marcus! Der Absturz beim Speichern auf Pixel-Geräten ist in Version 2.4.2 behoben — das Update rollt gerade aus. Danke für den Hinweis!”

also filed → Bug #213

🧑‍🤝‍🧑

Alpha / beta tester management

Soon

One shareable opt-in link per track. Nalya owns a Google Group per track and manages membership through the API — invites, removals, and a funnel showing who actually installed and gave feedback.

142

Invited

Joined

Installed

Feedback

Beta cohort crash-free rate: 99.4% · ready to promote → production

🔎

Automatic ASO

No SDKv1 · data in P2

The agent mines your reviews and competitor listings for the language users actually search, then drafts localized title, short and full descriptions — pushed to the store as a reviewable diff, never silently.

Wallpaper Studio — beautiful wallpapers for your phone

Wallpaper Studio — endless AI wallpapers, battery friendly

Evidence: “battery” appears in 41 reviews · localized & drafted · est. +6–9% conversion

🧪

A/B tests from automatic funnels

Soon

Connects to the PostHog or Firebase you already have (or provisions one for you). The agent reads your funnels, finds the drop-off, and proposes a concrete experiment — then creates it and reports significance.

App opened

100%

Browsed patterns

74%

Customized colors

28%

Set wallpaper

22%

💡 Suggested experiment: 46-point drop at color customization. Test showing 3 ready-made palettes before the full editor. Drafted as a 50/50 flag in PostHog.

📊

One growth dashboard

No SDKv1

Installs, ratings, store conversion, crash & ANR vitals, review sentiment, funnel health and every agent action — in one place, instead of four consoles. Plus a weekly digest that crowns the single task that drove the most growth — and names the frontrunner to beat it next week.

Store CVR

31%

▲ 2.1pt

Sentiment

+0.62

▲ improving

Agent actions

this week

🏆 This week's growth winner

Auto-replied to 38 reviewers stuck on “premium gone after reinstalling” with a one-tap fix.

+0.4★ rating recovered in 9 days — the single biggest gain, ahead of 22 other actions.

⏳ Next week's frontrunner: the 3-palette onboarding test — +6.1% activation and climbing.

🐛

Errors & bugs management

Soon

Crash and ANR clusters with stack traces straight from Play vitals — no SDK needed. The agent groups them, ranks by user impact, links the reviews that mention them, and drafts the GitHub issue for you.

NullPointerException · WallpaperRenderer.kt:1482.1% of usersissue drafted

ANR · input dispatch > 5s · PaletteEditor0.4%3 reviews linked

IOException · TextureCache.load0.1%watching

💰

Monetization autopilot

No SDKv1 · auto in P2

Revenue, MRR and churn straight from your Play financial exports — then the agent goes to work: purchasing-power-parity pricing per country, subscription offer tests, and win-back offers the moment a subscriber cancels. All through Play's official monetization APIs, and always supervised first — it's your revenue.

MRR$3,184 ▲ 9%

Premium · India — PPP suggestion₹449 → ₹199

Est. impact (based on 14 similar markets)units +31% · revenue +12%

Churn-save · 3 cancellations caught50% off × 2 months → 2 accepted

🧭

Deep onboarding for every capability

v1 · always

Each service comes with a guided, screenshot-by-screenshot setup that explains exactly which permission it needs and why — service accounts, Play Console steps, PostHog wiring, ads account linking. You always know what Nalya can touch, and nothing is requested before the capability that needs it.

STEP 3 OF 6

Grant “Releases” permission

In Play Console → Users & permissions, check Release to testing tracks. We need this to publish your betas — and nothing else yet.

✓ VERIFIED LIVE

Permission check passed

We validated the service account against your app and confirmed the exact scopes. You can revoke any of them at any time — capabilities degrade gracefully.

For developers

Or run the whole thing
from your editor.

Nalya ships an MCP server, so every capability is available to Claude Code, Cursor and any MCP-aware agent. Ask for your worst reviews, draft a reply, promote a release or pull revenue — in plain language, right where you already work. One key, scoped to a single app, revoke it anytime.

claude code — nalya mcp · reviews

▸ nalya: what are my worst reviews this week, and draft replies▋
→ list_reviews · 3 one-star reviews this week
→ summarize_reviews · top theme: crash on save (Pixel)
→ reply_to_review · 3 drafts in the reviewer's language
✓ Drafts ready — approve them here or in the console.

claude code — nalya mcp · releases

▸ nalya: promote the latest build to production▋
→ list_releases · v2.4.2 · open testing · 99.4% crash-free
→ promote_release · open testing → production · staged 25%
✨ release notes auto-generated from 14 commits · localized
✓ Live at 25% — I'll advance the rollout as vitals hold.

claude code — nalya mcp · aso

▸ nalya: tighten my store listing for search▋
→ get_aso_insights · "battery" in 41 reviews · missing from listing
→ compare_listing_variants · new title + descriptions · localized
✦ "endless AI wallpapers, battery friendly" · est. +6–9% conversion
✓ Diff ready for review — nothing goes live until you approve.

claude code — nalya mcp · a/b tests

▸ nalya: where am I losing users — propose an A/B test▋
→ get_acquisition · 46-pt drop at color customization
→ suggest_experiment · show 3 ready-made palettes first
→ implement_variant · builds the picker, ships it in v4.2
→ create_experiment · 50/50 flag live in PostHog
✓ Running — I'll report significance the moment it lands.

claude code — nalya mcp · a/b tests → build

▸ nalya: the picker test won — ship the winning variant▋
→ get_experiment · variant B won · +18% activation · p<0.05
→ suggest_options · 1 ship as default · 2 region-personalize · 3 add a 4th palette
→ draft_change · option 1 · here's exactly what I'll build:
"Make the 3-palette picker the first-run default. Move the legacy color sliders under Settings → Advanced. Keep the PostHog flag wired to the winner. Ship in v4.2."
✓ Pick an option and I'll build the winning variant right here — no copy-paste.

claude code — nalya mcp · reviews · reply

▸ nalya: reply to my latest 1★ — and help me say it right▋
→ get_review · 1★ from "Lucía" · es · "se cierra al guardar"
→ translate · "it closes the moment I hit save" · ES → EN
→ suggest_reply · 3 angles · apologize + fix ETA · ask device · offer beta
→ reply_to_review · drafted back in Spanish, matched to her tone
✓ Draft ready — approve to post, or tweak a line first.

claude code — nalya mcp · alert → fix

▸ nalya: the v5.2 rating alert — propose a fix I can ship▋
→ get_alert · v5.2 · "forced to sign in" · rating 4.6→4.2 · high
→ summarize_reviews · 218 reviewers want guest downloads back
→ propose_fix · put the login wall behind a remote flag · ship in v5.2.1
- if (!user) showLoginWall()
+ if (!user && flags.requireLogin) showLoginWall()
// flag defaults off · login stays optional · halt v5.2 · ship v5.2.1
✓ Fix ready to implement — approve and I'll open the PR + halt the rollout.

claude code — nalya mcp · reviews · auto-reply

▸ nalya: 9 users say premium's gone after reinstalling — handle it▋
→ list_reviews · 9 reviews · theme: "premium not restored on new device"
→ summarize_reviews · known cause · Play Store cache, not a billing bug
→ reply_to_review · auto-replied to all 9 with the step-by-step fix:
"Sorry for the hassle — this is a Play Store cache hiccup, not lost billing.
1 · Settings → Apps → Google Play Store → Clear storage
2 · Restart your device
3 · Reopen the app — your premium restores automatically."
✓ 9 replies posted · I'll watch for rating changes and follow up.

claude code — nalya mcp · reviews → repro

▸ nalya: any new bug hiding in this week's reports?▋
→ list_reviews · 12 mention "widget goes blank after the 5.3 update"
→ get_release · 5.3 · rewrote the widget refresh worker
→ write_repro · added test: widget renders after a WorkManager restart
→ run · ✗ fails — the worker never re-enqueues on app update
→ propose_fix · re-enqueue periodic work in onUpdate · patch ready
✓ Repro is red — confirmed bug. Approve and I'll ship the fix as 5.3.1.

Explore the MCP server →

The trust model

Autonomy is earned, not assumed.

Nobody hands their Play account to a robot on day one. Every capability starts in supervised mode — the agent drafts, you approve. After enough approvals without edits, you flip it to full-auto. Per capability. Reversible. Logged.

👀

1 · Supervised

Every action lands in your approval inbox with its evidence. Nothing touches the store without your tap.

📈

2 · Earning trust

The agent tracks its own accuracy: approvals without edits count toward graduation. You see the progress bar.

🤝

3 · Graduated

Flip the switch and that capability runs autonomously — within hard rails: never reply twice, never widen a rollout past a failing vitals gate, never spend past your cap.

🧾

Always · Audited

An immutable action log records everything the service ever did, exportable any time. Revoke a permission and the capability degrades to draft-mode instead of breaking.

💬Review repliesFULL AUTO

Graduated after 15/15 approvals without edits · 142 replies posted · 31 ratings revised upward

🚢Release publishingFULL AUTO

Graduated · 8 releases shipped · 1 rollout auto-halted on ANR spike (you slept through it — that was the point)

🔎ASO listing changesSUPERVISED

9/15 approvals without edits · graduation available in ~2 weeks at current pace

💰Price & offer changesSUPERVISED

5/15 approvals · revenue-touching actions stay supervised until you explicitly graduate them

📣Ad spend changesSUPERVISED

Money moves always require approval until you say otherwise · hard cap $20/day set by you

Built on official APIs

No scraping. No gray areas. No overpromising.

Everything Nalya does goes through documented Google and Apple APIs — which also means we're honest about where the platforms draw lines, and we design around them instead of pretending they don't exist.

Capability	Status	How it really works
Releases & staged rollouts	Fully automated	Play Publishing API: tracks, rollout percentages, localized release notes, AAB uploads.
Review replies	Fully automated	Play exposes the last ~7 days of reviews, so we poll continuously from the moment you connect — no review is ever missed. Replies capped at 350 chars by Google.
Crash / ANR vitals	Fully automated	Play Developer Reporting API — crash & ANR rates, stack traces, ~3 years of history, zero SDK.
Installs & ratings data	Fully automated	Ingested from your Play Console cloud exports — Google doesn't expose these in a live API; we wire the export bucket during onboarding.
Tester lists	Automated via Groups	Play's API only manages Google Groups (not raw email lists) — so Nalya owns one group per track and manages membership for you.
Listing copy updates	Fully automated	Title, descriptions and graphics per locale via the Publishing API — always shown as a diff first.
Pricing & subscriptions	Fully automated	IAP prices and subscription offers via Play's monetization endpoints, including per-region price conversion. Revenue ingested from your financial exports; cancellation signals via Play's real-time notifications. Price changes always start supervised.
iOS / App Store	Phase 3	App Store Connect API: TestFlight, full review history (better than Play!), product-page A/B test creation, analytics reports.

Ship the app.
We grow it.

Good morning, Adrian 👋

Building the app was the easy 20%.

A closed loop, not another dashboard.

It doesn't wait for you to notice.

Everything your launch needs. One connection.

Publishing with auto release notes

Review auto-replies

Alpha / beta tester management

Automatic ASO

A/B tests from automatic funnels

One growth dashboard

Errors & bugs management

Monetization autopilot

Deep onboarding for every capability

Or run the whole thing
from your editor.

Autonomy is earned, not assumed.

Your growth team,
in your pocket.

Connected in three steps. Growing the same day.

Connect your Play account

Optional: GitHub & analytics

Meet your agent

No scraping. No gray areas. No overpromising.

Live now — and what's next.

The MVP, free

On the roadmap

Deeper intelligence

Then everywhere

Free while we're in beta.

Your app deserves a
growth team that never sleeps.

Building the app was the easy 20%.

A closed loop, not another dashboard.

It doesn't wait for you to notice.

Everything your launch needs. One connection.

Publishing with auto release notes

Review auto-replies

Alpha / beta tester management

Automatic ASO

A/B tests from automatic funnels

One growth dashboard

Errors & bugs management

Monetization autopilot

Deep onboarding for every capability

Or run the whole thingfrom your editor.

Autonomy is earned, not assumed.

Your growth team,in your pocket.

Connected in three steps. Growing the same day.

Connect your Play account

Optional: GitHub & analytics

Meet your agent

No scraping. No gray areas. No overpromising.

Live now — and what's next.

The MVP, free

On the roadmap

Deeper intelligence

Then everywhere

Free while we're in beta.

Your app deserves agrowth team that never sleeps.

Or run the whole thing
from your editor.

Your growth team,
in your pocket.

Your app deserves a
growth team that never sleeps.