LT Google Review Booster

Neutral Google review requests for local businesses—without rating gates or incentives.

Comment

The problem I'm solving

Most "get more reviews" tools quietly do something they shouldn't: they screen customers by how happy they seem and funnel only the happy ones to Google. That's review gating, and in the US it's illegal — the FTC and the Consumer Review Fairness Act both come down hard on it. I wanted to build the opposite: a review-request tool that makes the honest path the easy path, and makes gating structurally impossible instead of just discouraging it.

What it does

LT Google Review Booster is a compliance-first Work & Productivity tool for local businesses. The owner never sees or touches the Google review link — LT keeps that destination hidden and only resolves it the moment a customer actually chooses to leave a review. Every customer gets the same neutral screen: "Review on Google" or "Send private feedback," with no star-rating question first, no sentiment filter, and nobody watching over their shoulder. Private feedback is a real, equal option, not a trap for burying bad reviews. The same rule holds across every channel — on-screen QR, printed QR, short links, receipts, NFC, and after-service follow-ups.

The working demo

There's a free, credential-free synthetic demo at https://lt-google-review-prod-dashboard-web.vercel.app/demo — no login, payment, API key, or account needed. You pick a synthetic location, activate or pause an on-screen QR placement, open the customer route, and see the two neutral choices. Tap "Review on Google" and it fails closed, because this demo has no live destination; the evidence panel then logs each step as a non-persisted event with no customer data.

How I built it

I'm a solo developer and I made all the product, compliance, and safety calls myself. I used OpenAI Codex with GPT-5.6 for the engineering: GPT-5.6 turned a dense compliance spec into working contracts, state machines, and tests, and Codex ran those changes across the whole monorepo and caught boundary bugs in review before I merged. The names Sol, Terra, Luna, and Spark in my commits are just labels for parallel tasks in the same Codex session — not different models. Under the hood it's a Node.js 24 / TypeScript monorepo with Next.js web apps, an Android/Expo shell, and shared policy and contract packages, where the anti-gating rules live in typed contracts backed by automated tests.

Being honest about scope

This is a runnable contest vertical slice, not a finished product. Live Google redirects, messaging, billing, production sign-in, and persistent storage are deliberately turned off until their privacy and provider controls are in place. The demo and this write-up only claim what actually runs today.

What's next

The same neutral, anti-gating workflow is designed to extend to printed materials, NFC, receipts, and completed-service triggers, and to more markets — reusing the exact same compliance rules so the "no gating" guarantee holds everywhere, not just on the web demo.

Built With

Share this project:

Updates

posted an update —

Demo video shipped — submission finalized

The short public demo video is now recorded and linked above. It's a screen capture of the credential-free build running on synthetic data: location selection, a scannable local QR route going active, the neutral two-choice customer screen (Review on Google / Send private feedback), the Google action failing closed with no live destination, and the audit-safe event log. Nothing is staged or generated — it's the real running demo, under three minutes.

That closes the last open item from the previous update. External providers remain disabled exactly as before: no live Google redirect, messaging, billing, or production data path. The build was engineered solo with Codex and GPT-5.6 — the /feedback session ID and the dated in-period commits are in the README.

Log in or sign up for Devpost to join the conversation.

posted an update —

Build progress — local demo and security baseline

LT Google Review Booster is now at a verified 46% implementation estimate. The credential-free build includes a deterministic synthetic merchant-to-customer workflow, a scannable local QR route, neutral customer choice, browser-local vault foundations, Android backup and permission hardening, server-only webhook boundaries, secret scanning, dependency checks, and CycloneDX SBOM evidence.

All external providers remain disabled: no live Google redirect, messaging, billing, production data path, or hackathon submission is claimed. Next: hosted judge access, separate-device evidence, and a short public demo video.

Log in or sign up for Devpost to join the conversation.

posted an update —

Runnable local demo: neutral choices and fail-closed states

A first interactive merchant-to-customer slice is now published in commit 6072d1d.

The local /demo route now lets a reviewer:

  • select a visibly synthetic organization and location;
  • activate or pause an on-screen placement preview;
  • open the neutral customer-choice surface;
  • verify that the Google action fails closed when no live destination exists;
  • acknowledge private feedback without collecting response or customer data; and
  • inspect event-state evidence labeled as non-production and non-persisted.

The flow uses progressive server-rendered GET states so it does not depend on client-side memory. Automated state tests, browser interaction checks, lint, typecheck, Next.js production builds, Android export, secret scanning, and the full repository gate pass.

An independent Codex 5.6 Terra audit then found three source-boundary gaps—repeated query stages, runtime-invalid actions, and missing synthetic organization context. Codex 5.6 Sol corrected and published them in efd5d4f.

Current weighted V1.0 estimate: 40%. This is still a draft, not a release or submission. The QR graphic is intentionally non-scannable; authentication, regional opaque routing, persistence, live redirects, hosted judge access, video, and final contest evidence remain open.

Log in or sign up for Devpost to join the conversation.

posted an update —

Build underway: verified foundation and parallel delivery

LT Google Review Booster is now 38% complete toward the production-ready V1.0 baseline. This is a weighted engineering estimate, not a release claim.

We replaced serial phases with prerequisite-driven build segments. Codex 5.6 Sol and Terra now work in the same Codex task using direct autonomous handoffs and exclusive file leases. That removes operator copy/paste and wake-up prompts while preserving source review, independent tests, and release gates.

Verified implementation now includes:

  • A credential-free Node.js 24 / TypeScript / pnpm monorepo and protected CI baseline
  • Isolated US/EU routing, migration, privacy, entitlement, and billing contracts
  • Fail-closed regional web authentication and an Android-first Expo shell
  • A field-minimized Google Places destination boundary
  • A neutral regional review-funnel presentation with 416 independent audit tests
  • Strict CSV mapping and row-normalization contracts, consent/suppression policy, completion-event verification, and request-template safety scanning
  • A full root gate that passes lint, typecheck, tests, Next.js production builds, Android export, and synthetic-fixture validation

The contest entry remains NO-GO for submission. Hosted services, live provider routes, billing activation, and production data are still disabled. The next product milestone is a coherent synthetic merchant-to-customer demo backed by the same compliance controls.

Code and evidence: https://github.com/gilberthlimocon/LT-Google-Review-Prod

Log in or sign up for Devpost to join the conversation.

posted an update —

Product specification and build plan complete

We finalized the product design, merchant workflows, Google-review compliance controls, US/EU architecture, database schema, global PPP pricing, subscription entitlements, and QA/release requirements.

OpenAI Codex 5.6 Sol and Terra helped organize the work into an 11-phase implementation tracker with clear ownership and validation gates.

Next milestone: Phase 1—building the runnable monorepo foundation, CI baseline, and synthetic demo environment. The project remains pre-build and is not yet ready for final submission.

Log in or sign up for Devpost to join the conversation.