FSRS Preset Advisor (Merger and Splitter)

AnkiWeb addon 487947984

Inspects, compares, and reorganizes FSRS presets across decks, suggesting mergeable groups by parameter similarity and supporting first-review deck splitting, merging back, and cleanup of unused advisor presets.
AI-generated summary; may contain mistakes.

fsrsstatisticsdeck-management

Open on AnkiWeb GitHub Find alternatives

AnkiWeb

Votes
0 (πŸ‘ 0 Β· πŸ‘Ž 0)
Updated
2026-07-06
Anki versions
2.1.1~
License
MIT (from the GitHub repository)
Virus scan
No ClamAV match A match does not prove malware. No match does not prove that the code is safe.
Languages
Python, JSON
Description language
en

Maintenance

active

  • Last update or commit: 2026-07-06 (118 days before the snapshot).
  • Test files: 8.
  • Commits in the last year: 22.
  • Issues: 0 open, 0 closed.
  • Pull requests: 0 open, 0 merged.

Will it work on my Anki?

Version branches
Min AnkiMax AnkiUpdated
2.1.1+2026-07-06
History across monthly snapshots

Loading…

Similar addons

Anki FSRS Preset Manager

Manages FSRS presets in a single table for editing retention, steps, and version, running optimize/evaluate, bulk-applying values, and setting deck-level overrides.

schedulingfsrsdeck-management

1 votes 21% approval πŸ‘ 1 Β· πŸ‘Ž 0 Anki 2.1.1~ Updated 2026-08-11
Advanced FSRS Simulator

Simplifies Anki's FSRS simulator by letting users simulate a deck, its subdecks, or a preset, inspect included cards and counts, adjust settings, and compare runs on one graph.

schedulingfsrsstatisticsdeck-management

1 votes 21% approval πŸ‘ 1 Β· πŸ‘Ž 0 Anki 2.1.2~ Updated 2026-09-23
SM-20 Custom Scheduler & Algorithm Arena

Brings experimental SM-2, SM-15, SM-19, SM-20, and FSRS v6 schedulers plus a weighted ensemble mode to Anki Desktop, with settings, prediction inspection, and optional optimization.

schedulingfsrsbeta

1 votes 21% approval πŸ‘ 1 Β· πŸ‘Ž 0 Anki 26.05~ Updated 2026-08-02
Leech Detector

Detects and classifies leeches using lapse metrics, adding Browse search filters, Card Info status, stats/plugin support, and a Tools summary with per-deck counts.

statisticssearchleech

2 votes 34% approval πŸ‘ 2 Β· πŸ‘Ž 0 ⭐ 2 Anki 2.1.1~ Updated 2026-03-04

Description

FSRS Preset Advisor

Always make a recent Anki backup before using this add-on. Some features can change deck assignments, create or delete presets, move cards between decks, and update FSRS preset parameters.

FSRS Preset Advisor is an Anki add-on that helps you inspect, compare, and reorganize FSRS presets across your decks.

It is intended for users who maintain several decks or subdecks and want help deciding which decks can share an FSRS preset, which presets look different enough to keep separate, and which advisor-created presets can be cleaned up later.

What it does

The add-on adds several actions to Anki's Tools menu:

  • FSRS Preset Proximity

    • Compares the FSRS parameters already saved on your existing presets.
    • Shows pairwise distances between presets.
    • Marks presets whose parameters are close enough to be considered similar.
  • FSRS Deck Proximity (Computed)

    • Computes FSRS parameters per deck scope, then compares decks by those computed parameters.
    • Defaults to leaf decks only.
    • Can optionally include middle and root decks.
    • Opens a grouping editor where you can compare current preset assignments with recommended similarity groups.
  • FSRS Mergeability (Evaluate)

    • Computes deck parameters and evaluates whether decks can share parameters using log loss comparisons.
    • Shows current preset groups, Mahalanobis-distance groups, and log-loss-based groups side by side.
    • Lets you apply one of the suggested splits to new or reused advisor presets.
  • FSRS Split Deck by First Review

    • Splits cards from a selected deck into Again, Hard, Good, and Easy target decks based on the card's first review rating.
    • Creates or reuses matching advisor presets for those target decks.
    • Leaves cards without a first review, or with an unexpected first rating, unchanged.
  • FSRS Merge Back First Review Split

    • Moves cards from the Again, Hard, Good, and Easy split decks back into the selected base deck.
    • Deletes empty split decks when possible.
  • FSRS Cleanup Empty Advisor Presets

    • Finds presets created by this add-on that are no longer assigned to any deck.
    • Asks for confirmation before deleting them.

How recommendations work

The main proximity workflow compares FSRS6 parameter vectors with a Mahalanobis distance model.

Only FSRS6-style parameter sets with 21 values are compared. Presets or computed deck parameters with a different shape are treated as not valid for the FSRS6 comparison.

The default shared-preset threshold is 3.8. In the deck-computed grouping editor, you can adjust the threshold with a slider and see the recommended groups update.

Recommended groups use a conservative complete-link style rule: a group is only recommended when every pair inside the group is below the selected threshold.

Changing presets and decks

Some screens are read-only comparison tools, while others can change your collection.

The add-on may:

  • assign decks to existing presets,
  • create advisor presets named like FSRS Preset Advisor : Group X,
  • create first-review split decks,
  • move cards between decks,
  • delete empty advisor-created presets after confirmation,
  • optimize FSRS parameters for advisor presets when you confirm that action.

Before using actions that move cards or reassign presets, make sure you have a recent Anki backup.

Stored files

The add-on stores small helper files in the active Anki profile folder when possible:

  • deck_params_cache.json

    • caches computed deck parameters and reuses them when the review count has not changed.
  • evaluate_logloss_cache.json

    • caches mergeability log-loss evaluations when the relevant review count has not changed.
  • deck_preset_backup.json

    • stores preset assignments so advisor-driven preset changes can be reverted from inside the add-on.

Older versions may have stored these files in the add-on folder. When legacy files are read successfully, the add-on migrates the data to the preferred profile-folder location.

Installation

If you have a packaged release, install the .ankiaddon file through Anki:

  1. Open Anki.
  2. Go to Tools -> Add-ons.
  3. Choose Install from file....
  4. Select dist/anki-fsrs-preset-advisor.ankiaddon.
  5. Restart Anki.

This add-on expects an Anki version with FSRS support and the optimizer APIs used by current Anki releases.

Development

Useful commands:

pytest -q
./copy_to_anki.sh

README

FSRS Preset Advisor

Always make a recent Anki backup before using this add-on. Some features can change deck assignments, create or delete presets, move cards between decks, and update FSRS preset parameters.

FSRS Preset Advisor is an Anki add-on that helps you inspect, compare, and reorganize FSRS presets across your decks.

It is intended for users who maintain several decks or subdecks and want help deciding which decks can share an FSRS preset, which presets look different enough to keep separate, and which advisor-created presets can be cleaned up later.

What it does

The add-on adds several actions to Anki's Tools menu:

  • FSRS Preset Proximity

    • Compares the FSRS parameters already saved on your existing presets.
    • Shows pairwise distances between presets.
    • Marks presets whose parameters are close enough to be considered similar.
  • FSRS Deck Proximity (Computed)

    • Computes FSRS parameters per deck scope, then compares decks by those computed parameters.
    • Defaults to leaf decks only.
    • Can optionally include middle and root decks.
    • Opens a grouping editor where you can compare current preset assignments with recommended similarity groups.
  • FSRS Mergeability (Evaluate)

    • Computes deck parameters and evaluates whether decks can share parameters using log loss comparisons.
    • Shows current preset groups, Mahalanobis-distance groups, and log-loss-based groups side by side.
    • Lets you apply one of the suggested splits to new or reused advisor presets.
  • FSRS Split Deck by First Review

    • Splits cards from a selected deck into Again, Hard, Good, and Easy target decks based on the card's first review rating.
    • Creates or reuses matching advisor presets for those target decks.
    • Leaves cards without a first review, or with an unexpected first rating, unchanged.
  • FSRS Merge Back First Review Split

    • Moves cards from the Again, Hard, Good, and Easy split decks back into the selected base deck.
    • Deletes empty split decks when possible.
  • FSRS Cleanup Empty Advisor Presets

    • Finds presets created by this add-on that are no longer assigned to any deck.
    • Asks for confirmation before deleting them.

How recommendations work

The main proximity workflow compares FSRS6 parameter vectors with a Mahalanobis distance model.

Only FSRS6-style parameter sets with 21 values are compared. Presets or computed deck parameters with a different shape are treated as not valid for the FSRS6 comparison.

The default shared-preset threshold is 3.8. In the deck-computed grouping editor, you can adjust the threshold with a slider and see the recommended groups update.

Recommended groups use a conservative complete-link style rule: a group is only recommended when every pair inside the group is below the selected threshold.

Changing presets and decks

Some screens are read-only comparison tools, while others can change your collection.

The add-on may:

  • assign decks to existing presets,
  • create advisor presets named like FSRS Preset Advisor : Group X,
  • create first-review split decks,
  • move cards between decks,
  • delete empty advisor-created presets after confirmation,
  • optimize FSRS parameters for advisor presets when you confirm that action.

Before using actions that move cards or reassign presets, make sure you have a recent Anki backup.

Stored files

The add-on stores small helper files in the active Anki profile folder when possible:

  • deck_params_cache.json

    • caches computed deck parameters and reuses them when the review count has not changed.
  • evaluate_logloss_cache.json

    • caches mergeability log-loss evaluations when the relevant review count has not changed.
  • deck_preset_backup.json

    • stores preset assignments so advisor-driven preset changes can be reverted from inside the add-on.

Older versions may have stored these files in the add-on folder. When legacy files are read successfully, the add-on migrates the data to the preferred profile-folder location.

Installation

If you have a packaged release, install the .ankiaddon file through Anki:

  1. Open Anki.
  2. Go to Tools -> Add-ons.
  3. Choose Install from file....
  4. Select dist/anki-fsrs-preset-advisor.ankiaddon.
  5. Restart Anki.

This add-on expects an Anki version with FSRS support and the optimizer APIs used by current Anki releases.

Development

Useful commands:

pytest -q
./copy_to_anki.sh
./package_ankiaddon.sh

Additional technical notes are in docs/README.md, docs/ARCHITECTURE.md, and docs/HANDOFF.md.

License

MIT License. See LICENSE.