Brierly | Forecasting Skill and Calibration Trainer Flutter App for Android and iOS With Admin icon

Brierly | Forecasting Skill and Calibration Trainer Flutter App for Android and iOS With Admin

Go to Website
Hot or Not
Brierly | Forecasting Skill and Calibration Trainer Flutter App for Android and iOS With Admin

Description

AI disclosure: AI tools were used to assist with parts of the code, documentation and graphics of this item. The author reviewed and tested the result. The product itself contains no AI features.

What it is

Brierly is a Flutter source code for a forecasting-skill and calibration trainer on Android and iOS, together with an operator console (MIKODES Admin Kit, Node.js server) to run it as a small product. For every public question the user types their own probability of Yes first, with the crowd’s estimate hidden, and seals the call. Only then does the app show the crowd. When the question is decided, the app scores the call with the Brier score and draws a calibration curve, so the user can see whether their 70 percent calls really come true about 70 percent of the time, and how they did against the crowd. All calls and scores stay on the phone. You rename it, add your store product ids and publish it under your own brand.

Who it is for

Developers and small studios who want a ready education utility for the stores, and teachers, newsrooms, communities and teams who want a judgement-training app with their own questions. Brierly is a learning tool: it has no wallet, no money, no trading and no links to any market. Questions come from a public data source the server reads (the keyless Manifold Markets public API by default, not affiliated) or from questions you write and resolve yourself in the console.

Demo Instructions

The mobile app has no user accounts, so it has no login. The operator console demo is read-only (every change is refused). To try it:

  1. Open the demo page below and download the demo Android app (arm64 APK), or build the real app for Android and iOS from the download with flutter run. The demo app reads live public questions from the demo server.
  2. Finish onboarding, open a question, drag the slider to your own probability and tap Seal my call. The crowd’s estimate appears only now.
  3. Open My calls to see the sealed call. Withdraw it if you like (a withdrawn call never counts).
  4. The Calibration tab fills after you have decided calls; a fresh install shows an honest empty state because questions settle only after their close date. Preview images use labelled sample history.
  5. Open the operator console with the read-only login below and look at the market data source, questions, paywall and features pages.

Live demo and docs

How you earn with it

  • Lifetime and yearly Pro unlock with restore, using the stores’ in-app purchase (store fees apply). Without Pro a user gets a number of calls per week and a window of history; both are settings.
  • Server-side purchase verification for Apple (signed transactions and Server API) and Google Play, off by default.
  • A revenue page in the console showing gross verified store transactions.
  • Paywall, product ids, limits, brand name, accent colour and feature flags are changed from the console without a new build.

Every money setting starts empty: no product ids ship, so the paywall reads “Purchases are not set up yet” until you add your own ids. Nothing in this item promises any income.

Third-party services and costs (not included in the price)

  • Store accounts and the lifetime and yearly store products (no product ids ship)
  • Release signing (Android keystore, iOS team and bundle id)
  • Hosting, domain and https for the server (the app works without it but then has no questions)
  • The public question source: the default is the Manifold Markets public API; read its current terms before you go live. A second provider can be switched on, but it only works where your server can reach it (some internet providers block it), so test from your host
  • Support, privacy and terms URLs (empty by default)
  • Store listing text and screenshots for your own app
  • Real-device testing before release

Core features

  • Blind Call: your own probability first, the crowd’s estimate only after you seal. The server never sends the estimate in a question list; it is served at the moment a call is sealed.
  • Brier score and Calibration Card: your mean Brier score, a reliability diagram, you versus the crowd on the same questions, a breakdown by category, and late calls (sealed within 24 hours of the close) scored apart.
  • My calls: open and decided calls, withdraw before the close; void questions and withdrawn calls never count.
  • Weekly featured set chosen by the operator, share card and CSV export.
  • Questions from a public source with quality filters: minimum number of participants, topic allow list, word deny list, a per-question hide and a horizon in days. Your own questions with an https resolution source, resolved Yes, No or Void with a reason that goes to the audit log.
  • Optional extras, off by default: an anonymous leaderboard (only aggregates, only after the user opts in) and opt-in usage counts.
  • Onboarding, light and dark themes, offline-first history and scores.

Workflow

Install, onboarding, open a question, seal your call, see the crowd, wait for the decision, read your score and calibration curve. The app reads its settings and questions from your server with ETag caching (rate limit 60 requests per minute per IP, health check included).

Admin: operator console

  • Brand (name, accent, support URL) and legal URLs
  • Paywall, limits and store product ids
  • Feature switches
  • Market data source and quality filters
  • Questions: hide, feature, add your own and resolve them
  • Announcement banner and force update screen
  • Opted-in installs (anonymous counts) and revenue
  • Team, audit log, 2FA and notifications (MIKODES Admin Kit 0.7.0)
  • Read-only demo mode

Security

  • Console with team roles, 2FA and an audit trail.
  • Purchase verification checks Apple signed transactions (JWS) and Google Play on the server.
  • Secrets live in server/.env (a .env.example without values is included).
  • Usage counts are opt-in and anonymous; forecasts never leave the phone.

Tech stack

App: Flutter 3.41, Dart 3.11, sqflite, in_app_purchase, share_plus. Server: Node.js with Hono and better-sqlite3.

Requirements

  • Flutter 3.41.4 or newer (Dart 3.11.1+); Android Studio for Android; Xcode for iOS (deployment target 16.0).
  • Node.js 26 LTS (22.12 or newer also works) for the server; Linux or macOS.
  • Apple Developer and Google Play developer accounts to publish and to create the store products.
  • Optional: App Store Server API key and Google Play service account for purchase verification.

External paid services are not included in the item price: developer accounts, store fees, hosting, a domain with https for the server, and any service you connect.

What is included

  • Flutter app source (Android and iOS)
  • Node.js server with the operator console
  • .env.example, third-party licence list and documentation

Good to know before you buy

  • The app shows questions only when a server is reachable; without one it shows an honest empty state and history still works. No sample questions are bundled.
  • Question text and estimates come from a public data source that you operate against. Its terms, uptime and content are outside this item; use the quality filters and the per-question hide.
  • Nothing was tested on a real device: the share sheet and store purchases need your own testing before release. The Dockerfile is untested. The Android release build is debug-signed until you add your own key.

Installation summary

1) Server: copy .env.example to .env, then npm ci, npm run build, npm start. 2) App: set your brand and server URL in app/lib/brand.dart (or鈥攄art-define=API_BASE_URL), set bundle id and signing, then flutter run or build. 3) Create the two store products and enter their ids in the console. Full steps are in the documentation.

Declarations

AI declaration: AI tools were used to assist with parts of the code, documentation and graphics of this item. The author reviewed and tested the result. The product itself contains no AI features.

Changelog

v1.0.0 (first marketplace release): Blind Call, Brier score and Calibration Card, My calls, weekly featured set, share card and CSV export, public question source with quality filters, custom questions with audited resolution, paywall with restore, announcement, force update, purchase verification, operator console.

Support

Support covers questions about the item and bugs in the included code. Help with your own store accounts, signing or custom changes is not part of item support. Contact: https://mikodes.com/support

Credits

Open-source packages are listed in THIRD-PARTY-LICENSES.md. Preview images are real screens; question lists are live public data and the personal history is labelled sample data; device frames are illustrations. Not affiliated with Apple, Google or any data provider; their names only describe the platforms or sources.

AuthorMikodes
Price$29
Sales0
Rating0 (0)
PublishedNov 19, 2026
Last UpdateNov 19, 2026
Compatible SoftwareFlutter 3.x
Demo Urlhttps://brierly-demo.mikodes.com
Mobile Files IncludedDart, YAML, JavaScript JSON, .kotlin, .plist, .swift, JavaScript JS, CSS, HTML

Categories

Tags

brier-scorecalibrationeducationforce-updateforecastingin-app-purchaseioslearningnodejspaywallprobabilityquizremote-configsubscriptionwhite-label

External listing. code.market does not sell this product and does not process payment for it. Pricing, payment, delivery, licensing, refunds, and support are handled by the third-party merchant on its own website. We may earn an affiliate commission from a referral. Details.