Nibomo: Open-Source AI Flashcards with FSRS Reviews

Nibomo is a flashcard and spaced-repetition app for learners who want to turn material into reviewable cards and return to it on a schedule. It combines AI-assisted card creation with FSRS reviews, offline mobile study, and optional cloud-connected features across Apple devices, Android, and the web.

From Flashcards Open Source App to Nibomo

The GitHub repository still carries the name Flashcards Open Source App, while the App Store and product site now call the app Nibomo. The repository README links to the same App Store product ID, so this is a rename of the project rather than a similarly named flashcard app.

The current listing is published by SAMO DANNI EOOD; Nibomo's documentation says it was built by Kirill Markin. That distinction matters when matching the store listing to the source repository, and the official README links the two directly.

Turn notes and photos into editable cards

Nibomo is designed for exam preparation, vocabulary, and technical terms or facts. The store description says you can turn notes and photos into flashcards, then group cards into decks and add tags so a growing collection stays searchable.

AI can help draft a card, improve its wording, or explain a difficult topic. Treat generated material as a first draft: review the answer for accuracy, keep only what you want to study, and make the prompt or source material specific enough to be useful.

AI creation is online; reviewing can be offline

AI features require an internet connection, according to the App Store listing. The project also describes optional AI chat and image or card assistance, so requests that use those features may send the prompt and relevant content to the configured hosted or external provider.

That is separate from ordinary flashcard review. You do not need to keep an AI conversation open to work through saved cards, and the mobile app can review them without a connection. Offline study is useful on a commute or during travel, but it does not mean that cloud sync or AI works offline.

FSRS spaces reviews around recall

Nibomo uses FSRS to decide which cards are due. Its feature page describes bringing difficult cards back sooner and waiting longer before showing familiar material again, so the review queue responds to your answers rather than repeating every card at the same interval.

The aim is to spend more of a study session on material that needs attention. FSRS schedules practice; it does not supply the underlying subject knowledge or guarantee an exam result. Good cards still need clear prompts and answers that you have checked.

Study offline and sync when you reconnect

The product supports reviewing saved cards on mobile without an internet connection. When the device reconnects, changes sync so you can continue on the web, iOS, or Android. This makes offline review practical without implying that every part of the hosted service is local.

Cloud sync is account-based: the current privacy policy says an email address is required for a hosted account, although limited guest or local features may be available without signing in. If you need a fully self-managed setup, the project documents a separate self-hosting path.

Keep decks portable and organized

Decks and tags give you a simple way to separate courses, languages, and topics. The app also shows review activity and study streaks, which can help you see whether a routine is forming without turning every session into a leaderboard.

Nibomo's feature page says exports can include cards, tags, and related media. Import and export are useful when you want a backup or need to move learning material; check the generated package and destination before sharing it, especially if cards contain personal notes or images.

Connect MCP and AI agents when useful

Nibomo exposes an MCP connector and Agent API for supported actions such as creating, improving, and organizing cards. That gives compatible AI clients a route into study material, but it is optional; the ordinary review flow does not require an agent connection.

An external AI client has its own operator and model provider, so information requested through that client can be processed under their terms as well as Nibomo's. Review which workspace and cards the client can access before connecting it.

MCP endpoint: https://mcp.nibomo.com/mcp

Open-source code and hosted privacy are separate

The app source is available under the MIT license, and the project documents self-hosting. Open code makes implementation paths inspectable, but it does not by itself mean that the hosted App Store experience stores no data or has the same operator as a self-hosted installation.

Nibomo's current privacy policy describes processing account and authentication details, cards and review history, uploaded files and media, support information, operational diagnostics, and usage events. The App Store privacy label also lists categories such as contact information, user content, identifiers, usage data, diagnostics, and purchases; Apple notes that the developer supplies these disclosures and they are not verified by Apple.

Optional hosted AI can involve prompts, card or workspace context, attachments, images, audio, transcripts, and model responses. The privacy policy says product analytics has a separate in-app control, while turning it off does not erase events already recorded. Read the current privacy policy for the full scope, and remember that self-hosted deployments are governed by their operator.

Install from the official App Store listing

For iPhone and iPad, use the official Nibomo: AI Flashcards listing. It is the current listing linked by the project README, under the updated Nibomo name; the App Store page currently shows an in-app-purchase option.

The listing requires iOS 18 or later and iPadOS 18 or later. Its compatibility section also lists Apple-silicon Macs running macOS 15 or later, and the README describes the Mac app as Mac Catalyst; however, the App Store page labels it designed for iPad and not verified for macOS. Check the live compatibility details before choosing it as a Mac desktop app.

The README also links to the official Google Play app for Android and the web review app. The hosted service, optional AI, and sync features have different network and account requirements from studying saved cards offline.

A practical fit for steady study

Nibomo is a good fit if you want one portable collection of cards for exams, vocabulary, or technical learning, and prefer scheduled reviews over deciding manually what to revisit. The combination of editable AI drafts, FSRS, offline mobile review, and export gives learners several ways to build a sustainable routine.

It may be less suitable if your requirement is that every feature works without an account or network: cloud sync and AI are online services, and the hosted privacy policy describes data processing beyond the MIT-licensed code. Start from the App Store listing, use offline review when disconnected, and choose the hosted or self-hosted model that fits your privacy needs.

Official App Store listing: https://apps.apple.com/us/app/nibomo-ai-flashcards/id6760538964
Android listing: https://play.google.com/store/apps/details?id=com.flashcardsopensourceapp.app
Web study app: https://app.nibomo.com/review
iPhone/iPad require iOS/iPadOS 18 or later. Apple's compatibility details list Apple-silicon Mac and macOS 15 or later, but the page also says the app is designed for iPad and not verified for macOS.