Repeto vs Noji (Formerly AnkiPro)

Choose Noji if you need a released web, iPhone/iPad, and Android flashcard app with rich cards, collaboration, a large vendor-reported library, and—on Premium—configurable review controls and bounded offline study. Consider Repeto after release if FSRS and an Anki .apkg exit path matter more—but Repeto is still pre-release and is not the safer choice for an immediate migration. Noji is the current name of the product formerly called AnkiPro; it is made by a separate company and is not official Anki.

Disclosure: Repeto publishes this comparison and has a commercial interest in the decision. We link to first-party Noji and Anki sources, state where Noji is stronger, and distinguish Repeto’s current pre-release codebase from public availability. Facts were checked on August 21, 2026.

Repeto vs Noji at a glance

Decision	Noji	Repeto
Product identity	Formerly called AnkiPro; Anki’s official FAQ says Noji is developed by a separate group and is unrelated to the Anki ecosystem	Repeto is a separate pre-release flashcard product; it does not claim to be an official Anki client
Public status	Released web, iPhone/iPad, and Android product	Repeto’s web app is in pre-release; an iPhone/iPad update is in preparation; Android is unreleased
Best fit	Learners who want a modern visual editor, collaboration, ready-made decks, mobile apps, and adjustable study modes	Learners evaluating a focused FSRS workflow with editable generation and Anki-package portability
Card creation	Manual front/back cards, reverse cards, fill-in-the-blank, rich formatting, and images; Premium adds image occlusion, generated text-to-speech, prompt-based AI generation, AI templates, and web-only AI card generation from PDFs	Current codebase supports manual editable cards and AI-generated proposals; individual material paths still need live verification
Review method	Noji’s live comparison page says it uses SM-2; current help documents three spaced-repetition presets, a non-SRS Fast Review mode, four SR response buttons, and Premium custom review steps/modifiers	Current codebase uses FSRS with Again, Hard, Good, and Easy ratings
Anki import	Imports `.apkg` and `.colpkg`; Noji says card progress can migrate when the source Anki export includes progress data, but it does not specify exact history or scheduling-field fidelity	Imports supported `.apkg` files; broad fidelity is not yet claimed
Export	Exports user-created decks as text-only CSV or Noji’s own OFC backup format; imported or duplicated decks cannot be exported through the documented flow, and no `.apkg` export is documented	Current codebase exports decks as `.apkg`; exact template, media, and history fidelity still needs corpus testing
Offline	Premium can download up to six decks of at most 200 cards each for offline study; only spaced-repetition decks are eligible, adding or editing cards is unavailable offline, and Fast Review is unsupported	Full offline behavior is not verified for Repeto’s pre-release surfaces
Sharing and library	Real-time collaboration, link sharing, and a vendor-reported library of more than 50,000 decks	Current codebase includes a free shared-deck market, but its public inventory and participation are unproven
Price	Free version with a 50-card daily study limit; Premium removes the limit and adds advanced features, but verify the current regional checkout price	Repeto 2.2 has no final public price that should be used for a purchase decision

The products overlap most strongly in spaced repetition, visual flashcards, Anki import, AI-assisted creation, and shared content. Noji wins on released availability and its vendor-documented library and collaboration features. Repeto’s clearest intended advantages are a named FSRS implementation and .apkg export.

Noji, AnkiPro, and official Anki are different products

Noji’s release notes say the product changed its name to Noji and that the underlying product remained the same. Before that change, it was known as AnkiPro.

That history does not make Noji part of Anki. The official Anki FAQ lists Anki, AnkiMobile, AnkiDroid, and AnkiWeb as the Anki ecosystem, then identifies “Anki Pro (now Noji)” as a separately developed product unrelated to that ecosystem. Noji can import Anki packages, but format compatibility is not corporate or account compatibility.

This distinction matters when searching an app store, buying a subscription, expecting AnkiWeb sync, or following an Anki support guide. A Noji account does not become an AnkiWeb account, and an official Anki workflow should not be assumed to work in Noji.

Where Noji is better than Repeto

Noji is a released cross-platform product with bounded offline review

Noji can be used on the web and through iPhone/iPad and Android apps. Premium offline mode lets a learner download decks before losing connectivity, continue spaced-repetition reviews, and sync progress after reconnecting.

The limits are material: Noji documents a maximum of six offline decks with 200 cards per deck, supports only spaced-repetition algorithms offline, and does not allow adding or editing cards without a connection. Even with those constraints, it is a real documented workflow. Repeto has no released surface with an equivalent tested offline matrix.

Noji has richer collaboration and a larger existing library

Noji supports real-time deck collaboration, direct sharing, public profiles, and publishing with average-rating, download, and studier statistics. Its help center says the public library contains more than 50,000 decks. Published decks are reviewed, and creators can update or unpublish them.

The collaboration paths have limits: imported decks cannot be shared, and individual subdecks can be shared only with their main deck. The publishing guide excludes imported decks and subdecks and says publishing a shared deck removes its sharing access.

Repeto’s current codebase includes a shared-deck market, but it has no comparable public inventory, participation history, or evidence of moderation under load. For a class or study group that wants to build a deck together today, Noji is the stronger option.

Noji exposes more mature card and study controls

Noji documents rich text, images, fill-in-the-blank cards, reverse cards, image occlusion, LaTeX, code blocks, text-to-speech, folders, subdecks, and several learning modes. Premium users can adjust review steps, Easy bonus, interval modifier, maximum interval, and new/maximum cards per day.

It also separates long-term spaced repetition from a non-SRS Fast Review mode—called General in older help articles—which cycles through all cards and repeats cards marked Not quite until they are marked Got it. Repeto’s narrower FSRS flow may be easier to understand, but it does not match Noji’s released editing and mode breadth.

Where Repeto’s current design may be better

These are current-codebase or target differences, not public-availability wins.

Repeto standardizes on FSRS

Noji’s live Anki comparison page says it uses SM-2, although its dated help articles do not name the underlying scheduler family. Those articles document three spaced-repetition presets—General Spaced Repetition, Study Languages, and Study Medical Materials—plus Premium manual controls. We found no first-party claim that Noji currently uses FSRS.

Repeto’s current codebase uses FSRS as the standard scheduler. That documented implementation choice may matter to learners who want an explicitly named FSRS foundation; Noji’s help center does not document FSRS, while its live comparison page attributes SM-2 to Noji. This does not prove either product produces better memory outcomes, and the two products still need a controlled queue test.

Repeto is building an Anki-package exit path

Noji documents .apkg and .colpkg import and says card progress can transfer when progress data is included in the source Anki export; it does not define the exact history or scheduling fields preserved. Its documented outward path is different: CSV exports text only and omits custom formatting, pictures, and image occlusion, while Noji’s own OFC backup format can include images, image occlusion, customization, and optional learning progress. Noji says only decks created by the user can be exported—not imported or duplicated decks.

Repeto’s current codebase exports a deck as .apkg and imports supported .apkg files. That is a more direct intended .apkg bridge to another Anki-compatible workflow. Repeto must still publish a compatibility corpus before calling a round trip lossless.

Repeto includes a conversation-to-card connection

Repeto’s repository contains an OAuth-based MCP connection for supported AI clients, with deck and card tools. The intended use is to create or revise study material without manually moving every answer from an AI conversation.

We could not verify an equivalent MCP connection in Noji’s official product documentation. Absence from documentation is not proof that Noji will never offer one, and Repeto must verify client-specific setup in its released service before marketing this as generally available.

Where Repeto is worse or less proven

Noji is released on the web, iPhone/iPad, and Android; Repeto is not generally released under its current product identity.
Noji is a released product with a vendor-reported 50,000-plus-deck library and documented collaboration workflows. Repeto has no established add-on ecosystem, collaboration history, or shared-market network effect.
Noji documents image occlusion, rich formatting, LaTeX, generated text-to-speech, three SR presets, Premium custom review intervals, and a non-SRS Fast Review mode. Repeto has not demonstrated comparable public depth.
Noji has a real, if limited, offline review mode. Repeto must not promise full offline creation, editing, import, review, or later sync.
Repeto’s .apkg support lacks a published corpus covering custom templates, media, cloze behavior, modern variants, and review history.
Repeto has no released Android app and no public evidence that the iPhone/iPad update participates reliably in its current web/API sync system.

Noji has its own portability weakness: its current export documentation lists CSV and Noji’s own OFC backup format, not .apkg or .colpkg. It says OFC can include images, image occlusion, customization, and optional learning progress, but it does not document OFC compatibility with Anki or another app.

Card creation: mature visual tools or a focused proposal queue

Noji lets a learner create richly formatted cards manually, reverse batches, use fill-in-the-blank and image occlusion, use generated text-to-speech, and organize subdecks. Premium adds AI card generation, web-only PDF import, AI templates, and enhanced text-to-speech. Cards generated from PDFs can be edited before saving and study; Noji’s current help center says users cannot attach their own audio files.

Repeto’s pre-release codebase supports manual cards and editable AI-generated proposals inside a deck. Its intended workflow emphasizes reviewing proposed cards before they become permanent study items. Noji also supports editing generated cards, so Repeto should not present editability as unique.

Test the format you need. A medical learner using image occlusion has a different decision from a language learner who needs generated text-to-speech, or a researcher who wants cards grounded in external sources.

Reviewing: presets and manual controls versus FSRS

In spaced-repetition mode, Noji offers four response buttons—Again, Hard, Good, and Easy—and uses them to schedule the next review. Premium advanced settings expose review steps, Easy bonus, interval modifier, maximum interval, and daily card limits. Noji says switching between spaced repetition and Fast Review resets study progress; returning from Custom settings to a preset discards the custom settings.

Repeto’s current codebase also uses four ratings but lets FSRS model the card’s memory state. The presence of identical button labels does not imply identical intervals. Noji’s settings may suit a learner who wants direct interval controls; Repeto’s model may suit someone who prefers an explicit FSRS foundation.

Compare the same answer sequence for at least seven days. Record intervals and daily workload rather than deciding from the settings screen.

Anki import and export: direction matters

Noji documents .apkg and .colpkg import and says card progress can transfer when the source Anki export includes progress data; it does not define whether full review history or every scheduling field survives. A successful import still needs inspection: confirm note/card count, media, cloze behavior, image occlusion, tags, templates, and review state.

Noji’s documented export flow does not reverse that package path: it lists CSV and OFC, not .apkg or .colpkg. CSV is text-only and omits custom formatting, pictures, and image occlusion, although Noji says Cloze is supported; OFC can include images, image occlusion, customization, and optional learning progress. Only decks created by the user are exportable through this flow; imported and duplicated decks are excluded.

Repeto’s current codebase uses supported .apkg in both directions. Keep the original source package and test a representative deck before relying on either product as the only copy.

Platforms, offline use, and price

Noji has public web, iPhone/iPad, and Android surfaces. Its free version can be used indefinitely with essential features but limits study to 50 cards per day. Premium removes that limit and adds richer editing, images, image occlusion, configurable algorithms, generated text-to-speech, AI generation, web-only PDF imports, and offline learning. Noji’s help pages reviewed for this article do not provide one stable universal price, so verify the amount, billing period, trial, and renewal terms in your regional checkout.

Repeto cannot make a comparable released offer yet. The web app is pre-release, its iPhone/iPad update is in preparation, Android is unreleased, full offline behavior is unverified, and the final 2.2 price should be checked after launch.

Run the same 50-minute trial

Run the Repeto side only with legitimate pre-release access or after release.

Create ten manual cards in each product: two basic, two reverse, two cloze/fill-in-the-blank, two with images, and two with formatted explanations.
Use the same five-page PDF to request 20 generated cards, running the Noji side on the web; Noji documents a 50 MB limit per PDF. Count unsupported answers, duplicates, overloaded questions, and cleanup minutes.
If image occlusion matters, build one card from the same labeled diagram and record every manual step.
In each product’s spaced-repetition mode, review ten cards with the same Again/Hard/Good/Easy sequence and record the next intervals.
Return for seven days and note due-card count, missed-day behavior, and time to complete the queue.
Import the same representative .apkg into both products and inspect text, media, templates, tags, card count, and review history.
Export the imported and one newly created deck. Record which formats are offered and what a clean Anki profile can read.
Download the same-sized deck for offline use, disconnect, review ten cards, try one edit, reconnect, and verify progress elsewhere.
Share a copy with a study partner and record whether both people can edit, how conflicts behave, and whether the recipient owns an independent copy.

Do not use a simple feature score. Decide from accurate cards after review, repair time, seven-day workload, offline reliability, collaboration behavior, and the data that survives an exit.

Which should you choose?

Choose Noji if you need a released cross-platform app now, want rich visual cards and configurable intervals, need real-time collaboration or a large ready-made library, or can work within its bounded Premium offline mode. It is also the lower-risk choice for immediate Android use.

Consider Repeto after release if FSRS as the standard scheduler and .apkg export are more important than Noji’s mature editor, established library, and study modes. Preserve Anki backups until Repeto’s import/export test corpus and released sync behavior are proven.

Do not choose Noji because its former name contained “Anki,” and do not choose Repeto because code exists in a repository. Choose from the released workflow, scheduler behavior, and exit format you can verify yourself.

Official sources

Checked August 21, 2026. Noji facts come from its product/help pages; the identity clarification comes from the official Anki FAQ; Repeto availability follows the project’s release ledger.

Repeto vs Noji: which spaced-repetition app fits your study system?