Rekeeply Specification on Rekeeply Spec

book

Mon, 01 Jan 0001 00:00:00 +0000

book#

Fields:

author (text) - primary author of the book.
published_year (integer) - publication year of the edition used.
original_title (text) - title in the original language.
co_authors (text or long_text) - additional authors/co-authors.
language (text) - language of the specific edition/content.
original_language (text) - original language of the work.
publisher (text) - publishing organization.
publication_place (text) - city/place of publication.
edition (text) - edition label or number.
isbn (text) - ISBN identifier.
pages_total (integer) - total number of pages.
translator (text) - translator name(s).
volume (text or integer) - volume number/label for multi-volume books.
series (text) - book series title.
reference_scope (text) - scope note for book/reference entries.

article

Mon, 01 Jan 0001 00:00:00 +0000

article#

Fields:

author (text) - primary author of the article.
published_year (integer) - publication year.
journal_or_source (text) - publication source (journal/site/platform).
publication_date (date) - exact publication date.
doi (text) - DOI identifier.
volume (text or integer) - volume number/label.
issue (text or integer) - issue number/label.
pages (text) - page range (e.g., 97-111).
language (text) - article language.
original_title (text) - title in original language.
co_authors (text or long_text) - additional authors/co-authors.
peer_reviewed (boolean) - peer-reviewed marker.
publisher (text) - publisher/organization.
publication_place (text) - publication place.

Subtype notes:

Content Typing

Mon, 01 Jan 0001 00:00:00 +0000

Content Typing#

Baseline Rules#

content_type is mandatory; entries MUST NOT be created without a type.
content_subtype is optional.
Manual entry creation MUST start with selecting content_type.
After type selection, subtype and other fields may be filled later.

Default and Recommended Set#

The following type/subtype set is available by default and is recommended based on practical personal usage patterns.

`book`#

Purpose: long-form written works and references.

Fields: see Type-Specific Fields.

fiction - fictional books.
non_fiction - non-fictional books.
reference - lookup/reference-oriented books.

`article`#

Purpose: shorter written materials from web or publication sources.

Raw Model

Mon, 01 Jan 0001 00:00:00 +0000

Raw Model#

Purpose#

Raw is a separate domain because some user data must be stored in unstructured form before later processing.
Raw enables quick capture without mandatory classification or immediate placement into Library.

Contract#

Core implementations MUST include Raw as an account-scoped storage domain for unstructured capture.
Raw MUST support quick save without selecting an active library.
Raw MUST support review and deferred processing of captured items.
Raw items MUST be eligible for explicit later promotion into Library.

Supported Payload Types#

Raw MUST support text, link, and mixed payload types.
Raw MAY support file_ref as reference-only payload.
image, video, and audio payload types are out of the mandatory baseline scope.

Raw Item Storage Model#

Storage SHOULD use typed-columns-first design for hot read paths.
Implementations MUST avoid JSON-only storage for list/filter hot paths.
Each RawItem MUST include: id, account_id, payload_type, status, created_at, updated_at.
Implementations SHOULD include fast list fields such as title_preview and source/type/status columns.

Capture Inputs#

Raw MUST accept capture from manual input and clipboard/paste flows.
Raw SHOULD accept external share/import flows when available in client/runtime.
Capture MUST remain offline-capable and MUST NOT require network availability.

Metadata Enrichment#

For link payloads, metadata enrichment MAY run asynchronously after capture.
Metadata fetch failures MUST NOT block capture completion.
When metadata is unavailable, implementations MUST provide deterministic fallback preview.

Lifecycle and Operations#

Implementations MUST support item states equivalent to new | triaged | promoted | archived.
Implementations MUST support listing and status transitions for triage.
Implementations MUST support explicit promote from Raw to Library.

Promote Flow Requirements#

promote MUST validate that target_library_id belongs to the same account_id.
promote MUST be idempotent for the same RawItem and MUST NOT create duplicates on retry.
promote MUST persist linkage to created library entity (promoted_content_id or equivalent origin link).

Promote Mapping Rules#

Raw:text

Content Fields

Mon, 01 Jan 0001 00:00:00 +0000

Content Fields#

Field Classes#

system fields are managed by core and MUST NOT be manually edited by users.
user_editable fields are visible and editable by users.
computed fields are derived by core and SHOULD be treated as read-only.

Common Field Baseline#

`system`#

id - immutable entry identifier.
library_id - owning library identifier.
created_at - creation timestamp.
updated_at - last update timestamp.
status_changed_at - timestamp of last status change.

`user_editable`#

content_type - required type key.
content_subtype - optional subtype key.
title - entry title.
status - inbox | todo | in_progress | skimmed | done | dropped.
priority - low | medium | high.
rating - optional integer rating in range 1..10.
recommendation - optional recommendation marker: recommended | neutral | not_recommended.
favorite - independent favorite flag.
archived - independent archive flag.
parent_content_id - optional parent relation.
source_url - optional source link.
summary - optional short description.

`computed`#

search_text - normalized text index for fast search.
last_activity_at - derived last activity timestamp.
is_completed - derived completion flag from status.

Attachment Model#

Content entries MAY have multiple linked attachments.
Attachment links are modeled as related entities, not as one-off fixed fields.
Attachment baseline fields: id, content_id, uri, name (optional), mime (optional), size (optional), created_at.
Attachment export/import baseline dataset file: attachments.ndjson.

Rules#

Clients MUST enforce non-editability for system and computed fields.
Core MUST maintain consistency of all computed fields.
On creation, status SHOULD be auto-set to inbox.
On creation, priority SHOULD be empty by default unless explicitly provided.
Status changes SHOULD update status_changed_at and SHOULD NOT update updated_at.
Favorite flag changes SHOULD NOT update updated_at.
Archived flag changes SHOULD NOT update updated_at.
If rating is provided, value MUST be in range 1..10.
User-editable fields other than content_type and status are optional at creation time and may be filled later.
Type-specific fields are non-mandatory by baseline and become required only when user-defined schema marks them required.
Implementations MAY add extra fields in any class as implementation-defined extensions.

game

Mon, 01 Jan 0001 00:00:00 +0000

game#

Fields:

platform (text) - primary runtime platform/environment (e.g., PC, PlayStation, Xbox, Nintendo, board).
release_year (integer) - initial release year.
developer (text) - game developer/studio.
publisher (text) - game publisher.
distribution_source (text or enum) - source/store where game is available or owned (e.g., Steam, Epic Games Store, PS Store, Xbox Store, Nintendo eShop).
genre (text or multi_enum) - genre classification (e.g., RPG, strategy, puzzle).
mode (enum or multi_enum) - gameplay mode (single-player, multiplayer, co-op).
playtime_hours (decimal) - estimated or tracked playtime in hours.
players_count (text or integer) - supported/typical number of players.
edition (text) - game edition/version label.
region (text) - regional version/availability scope.

Field Types

Mon, 01 Jan 0001 00:00:00 +0000

Field Types#

Scope#

This section defines allowed field_type values for user-defined fields.
For all field types, required is configurable (true | false).
Default values are configurable when supported by the field type.

Allowed Field Types#

text - short text value.
- Validation: min_length (optional), max_length (optional), pattern (optional), trim_on_save (optional).
- Default: default_value (optional string).
long_text - long multiline text value.
- Validation: min_length (optional), max_length (optional), trim_on_save (optional).
- Default: default_value (optional string).
boolean - true/false value.

screen

Mon, 01 Jan 0001 00:00:00 +0000

screen#

Fields:

release_year (integer) - release year.
format (text or enum) - screen format context (e.g., film/series/anime/animation).
original_title (text) - title in original language.
country (text) - production country.
director (text) - director.
writers (text or long_text) - writers/screenwriters.
cast (long_text) - main cast.
studio_or_network (text) - studio/network/platform.
season_number (integer) - season number (especially for screen/episode).
episode_number (integer) - episode number (for screen/episode).
duration_minutes (integer) - runtime in minutes.
release_date (date) - exact release date.
age_rating (text) - age rating/classification.
imdb_id (text) - IMDb identifier.
franchise (text) - franchise/universe name.

Subtype notes:

video

Mon, 01 Jan 0001 00:00:00 +0000

video#

Fields:

creator_or_channel (text) - primary creator, author, or channel.
release_year (integer) - release/publication year.
original_title (text) - title in original language.
duration_minutes (integer) - video duration in minutes.
platform (text) - hosting platform (e.g., YouTube, Vimeo, Twitch).
series_name (text) - series/show name if video belongs to a series.
episode_label (text) - episode/part label.
language (text) - spoken/content language.
subtitles (multi_enum) - subtitle languages/availability.
published_at (datetime) - exact publication timestamp.
age_rating (text) - age suitability marker.
view_count_snapshot (integer) - optional captured view count.

Subtype notes:

quote

Mon, 01 Jan 0001 00:00:00 +0000

quote#

Fields:

quote_text (long_text) - captured quote text or idea statement.
quoted_author (text) - person/source author of the quote.
source_title (text) - title of the source work.
source_locator (text) - source pointer (page, timestamp, chapter, URL fragment).
language (text) - quote language.
original_text (long_text) - original-language quote if translation is used.
context_note (long_text) - context/interpretation note.
quote_date (date) - date of saying/publication when known.

Subtype notes:

quote/text usually uses source_locator and quoted_author.
quote/dialogue usually uses source_title and context notes.
quote/idea usually uses context_note with concise source linkage.

Tags#

Position in Model#

Tag is a dedicated library-scoped entity and is NOT a field_type.
Content entries are linked to tags via many-to-many association.

Content Association#

Content-tag relation MUST support multiple tags per content entry.
Core MUST support assigning existing tags to content entries and unassigning tags from content entries.
Core SHOULD support creating tags and assigning them in one logical operation.

Editing Rules#

Renaming a tag updates label without requiring content rewrites.
Removing a tag MUST remove links from content entries but MUST NOT remove content.
Merging tags MAY be supported and SHOULD preserve content links.

Import/Export and Sync#

Tag taxonomy MUST travel with library export/import.
Tag data and tag-content links MUST be synchronized in sync-enabled mode.

Library Actions

Mon, 01 Jan 0001 00:00:00 +0000

Library Actions#

Scope#

These actions define baseline operations for library entries and library tags.

Entry Actions#

create
- Starts with selecting content_type.
- Core auto-fills system fields (id, library_id, created_at, updated_at).
- Core sets default status=inbox and default empty priority.
edit
- Updates user-editable fields.
- Updates updated_at.
set-subtype
- Sets or clears content_subtype.
- MUST NOT change content_type.
- Updates updated_at.
set-status
- Sets status from allowed status set.
- Free transition model (no hard state machine).
- MUST update status_changed_at.
- MUST NOT update updated_at.
set-priority

person

Mon, 01 Jan 0001 00:00:00 +0000

person#

Fields:

name (text) - primary person name.
full_name (text) - full legal/full public name.
aliases (long_text) - alternative names/aliases.
born_year (integer) - birth year.
died_year (integer) - death year when applicable.
nationality (text) - nationality/country identity marker.
roles (multi_enum) - roles/professions.
known_for (long_text) - short note about what person is known for.
official_site (url) - official website/profile URL.
external_ids (json) - external identifiers (e.g., ORCID, Wikidata, IMDb).

Subtype notes:

resource

Mon, 01 Jan 0001 00:00:00 +0000

resource#

Fields:

resource_name (text) - primary resource title/name.
resource_kind (text or enum) - resource kind descriptor inside subtype context.
provider (text) - provider/owner organization.
official_url (url) - primary resource URL.
access_model (enum) - access mode (free/paid/freemium/private).
license (text) - license/usage model.
pricing_note (text) - pricing model summary.
api_available (boolean) - API availability marker.
repo_host (text) - host platform for repositories (e.g., GitHub, GitLab).
language (text) - primary language of content/interface.
last_checked_at (datetime) - last verification/check timestamp.

Subtype notes:

Schema Actions

Mon, 01 Jan 0001 00:00:00 +0000

Schema Actions#

Scope#

These actions define lifecycle operations for library-scoped schema entities: custom types, custom subtypes, and custom fields.
Schema entities use internal stable key values that are unique within one library scope.

Type Actions#

create-type
- Creates a custom content type in current library.
- Internal key MUST be unique within library.
edit-type
- Updates editable type properties.
- Internal key MUST remain immutable.
archive-type / unarchive-type
- Disables/enables type for active usage in new entries.
- Existing entries using this type remain valid.
delete-type

music

Mon, 01 Jan 0001 00:00:00 +0000

music#

Fields:

artist (text) - primary artist/composer.
original_title (text) - original track/release title.
album_title (text) - parent album title.
release_year (integer) - release year.
duration_seconds (duration) - track/runtime duration.
genre (text or multi_enum) - music genre.
label (text) - record label.
isrc (text) - ISRC identifier.
platform (text) - listening platform/service.
track_number (integer) - track order in album/EP.
disc_number (integer) - disc number for multi-disc releases.

Subtype notes:

music/track usually uses duration_seconds, album_title, track_number.
music/melody usually uses concise metadata with optional source linkage.
music/album and music/ep usually use release_year, label, and high-level notes.
music/playlist usually uses platform/source and curation notes.

audio

Mon, 01 Jan 0001 00:00:00 +0000

audio#

Fields:

speaker_or_author (text) - primary speaker/author/host.
original_title (text) - title in original language.
release_year (integer) - release/publication year.
duration_seconds (duration) - audio duration.
language (text) - spoken language.
publisher_or_show (text) - publisher/show/network name.
episode_label (text) - episode/chapter label.
narrator (text) - narrator/voice performer.
platform (text) - distribution/listening platform.
isbn_or_external_id (text) - external identifier when applicable.

Subtype notes:

audio/audiobook usually uses narrator, publisher_or_show, isbn_or_external_id.
audio/podcast_episode usually uses episode_label, publisher_or_show, platform.
audio/podcast_show usually uses show-level publisher/platform metadata.
audio/lecture_audio and audio/interview_audio usually use speaker_or_author, duration_seconds, language.
audio/voice_note usually uses minimal metadata and creation context notes.

Presets

Mon, 01 Jan 0001 00:00:00 +0000

Presets#

Contract#

Preset is a saved search definition in library scope.
Preset stores search conditions only.
Preset MUST NOT store sorting or displayed field layout.
Preset MAY be nested using parent_preset_id.
Preset MAY define inherit_query (true|false) for parent query inheritance.
Preset MAY include visual metadata (color, icon).

Operations#

create-preset
update-preset
rename-preset
delete-preset
run-preset

Hierarchy Rules#

Preset hierarchy cycles are forbidden.
If inherit_query=true, resulting query is combined with parent query using baseline AND composition.
If inherit_query=false, parent relation is organizational only.
On parent preset deletion, child presets SHOULD be moved to root level.

User Control#

User can define and edit preset names.
User can define preset order.

Collections

Mon, 01 Jan 0001 00:00:00 +0000

Collections#

Contract#

Collection is a manual library-scoped grouping of content entries.
Collection does NOT store search conditions.
One content entry MAY belong to multiple collections.

Collection Fields#

id (system) - immutable collection identifier.
library_id (system) - owning library scope.
name (user_editable) - collection name.
description (user_editable, optional) - collection description.
color (user_editable, optional) - visual color marker.
icon (user_editable, optional) - visual icon marker.
parent_collection_id (user_editable, optional) - parent collection for hierarchy.
created_at (system) - creation timestamp.
updated_at (system) - update timestamp.

Operations#

create-collection
rename-collection
delete-collection
add-to-collection
remove-from-collection
reorder-collections

Hierarchy Rules#

Parent collection MAY have multiple child collections.
One collection MAY have at most one parent.
Collection hierarchy cycles are forbidden.

misc

Mon, 01 Jan 0001 00:00:00 +0000

misc#

Fields:

kind_note (text) - short note describing what this entry represents.
external_id (text) - optional external identifier.
extra_metadata (json) - advanced custom metadata object.

Subtype notes:

misc has no default subtypes.

Execution Model

Mon, 01 Jan 0001 00:00:00 +0000

Execution Model#

Execution Modes#

embedded (default): core runs in the same process as a client application for minimum latency.
service: core runs as a separate local process and is accessed by clients via local IPC.
remote: client connects to a remote core backend over a secure network channel.

Common Rules#

same core domain logic;
authentication is mandatory;
one active session is bound to one account;
library-scoped operations work with only one active library;
clients use the core contract, not direct DB file access.

Session Lifecycle#

init: initial local vault setup and first library creation;
unlock: user authentication and opening a working session;
active: command execution in active session context (library-scoped or account-scoped);
switch-library: explicit active library switch inside a session;
auto-lock: automatic lock on inactivity timeout;
manual lock/logout: explicit lock or session termination.

Sync Model (High-Level)#

local DB is the default source of user operations;
synchronization is asynchronous and does not block user actions;
synchronization supports both library-scoped data and account-scoped data;
in offline mode, core continues to operate and sync is deferred;
conflicts are resolved by a deterministic policy (defined in a separate section);
sync data transfer must use a secure channel.

Search Behavior

Mon, 01 Jan 0001 00:00:00 +0000

Search Behavior#

Core Rule#

Search and Preset MUST use the same query engine and semantics.

Search vs Preset#

Search is ephemeral by default and is not persisted unless explicitly saved.
Preset is an explicitly persisted search definition.
When current search is changed or library session is restarted, previous ephemeral search result is discarded.
Only explicitly saved presets remain available across sessions.

Searchable Dimensions#

Search MUST support structured conditions by status, priority, rating, recommendation, favorite, archived, content_type, and content_subtype.

Authentication Boundary

Mon, 01 Jan 0001 00:00:00 +0000

Authentication Boundary#

authentication rules are defined by core;
credential validation and session issuance are performed by core;
clients only submit auth data through the core contract and store client-side session artifacts;
data access without an active valid session is forbidden.

Library Lifecycle

Mon, 01 Jan 0001 00:00:00 +0000

Library Lifecycle#

Library Attributes#

id - unique immutable identifier.
name - editable library name.
icon - editable visual marker.

Lifecycle Operations#

create-library
- Creates new library with id, name, and icon.
rename-library
- Updates name only.
- id remains immutable.
update-library-icon
- Updates icon only.
archive-library / unarchive-library
- Archived library MUST NOT be selected as active until restored.
delete-library
- Hard delete requires explicit confirmation.
- Before delete, system SHOULD show related data volume summary.
- If deleted library is active:
  - switch active library to fallback when available;
  - otherwise clear active library until next selection/creation.

Invariants#

Lifecycle operations are account-scoped.
Cross-account lifecycle operations are forbidden.
Library id is stable and never reused in same account scope.

Account Model

Mon, 01 Jan 0001 00:00:00 +0000

Account Model#

Contract#

Core implementations MUST include Account as a first-class domain entity.
One account MUST own zero or more libraries; each library MUST belong to exactly one account.
One active session MUST be bound to exactly one account.
Library-scoped operations in one active session MUST operate on exactly one active library.

Account Attributes#

id MUST exist and MUST be immutable.
status MUST exist and represent lock boundary (locked/unlocked or equivalent).
security_policy_ref MUST exist and MUST NOT contain plaintext secrets.
default_library_id MAY exist; when present it MUST reference a library owned by the same account.
Additional metadata MAY exist and is implementation-defined.

Lifecycle#

init MUST create exactly one local account in a new vault.
Repeated init on an initialized vault MUST be rejected unless an explicit reset/force flow is used.
Runtime state machine MUST be equivalent to uninitialized -> locked -> unlocked -> locked.

Session and Access Rules#

After unlock, session context MUST be bound to one account_id.
Raw-scoped operations MAY run without selecting an active library.
Cross-account operations in one active session MUST be rejected.
Write operations on domain entities MUST be rejected while session is locked.
Export/import operations MUST run in the active account context.

Non-Goals#

No email/username/avatar requirements.
No OAuth or social identity requirements.
No client/device UI settings in core account schema.

Library Transfer Format

Mon, 01 Jan 0001 00:00:00 +0000

Library Transfer Format#

Baseline#

Library export/import MUST use the *.rkl file extension.
*.rkl container MUST use tar.zst.
Text datasets inside *.rkl MUST use UTF-8 encoding.
Baseline library export scope includes library-scoped data only.
Raw data is excluded from baseline library export/import scope.
Transfer package MUST include manifest metadata, format version, and integrity data.

Package Structure Baseline#

manifest.json (required)
library.json (required)
taxonomy.json (required)
content.ndjson (required)
tags.json (required)
tag_links.ndjson (required)
attachments.ndjson (required)
progress.ndjson (optional)
notes.ndjson (optional)

Manifest Baseline#

manifest.json MUST include: format, version, encoding, container, exported_at, library, datasets, integrity.
format MUST be rkl.
encoding MUST be utf-8.
container MUST be tar.zst.
library MUST include id and name.
datasets MUST include dataset path and checksum for each included dataset.
integrity MUST include checksum algorithm and manifest checksum.

Dataset Baselines#

library.json MUST include id, name, icon.
taxonomy.json MUST include type/subtype taxonomy and field schema definitions.
content.ndjson rows MUST include id, type, status, created_at, updated_at, status_changed_at.
tags.json MUST include tag metadata.
tag_links.ndjson rows MUST include content_id, tag_id.

Taxonomy Transfer Rule#

Taxonomy MUST travel with library during export/import.
Taxonomy scope includes: content types, content subtypes, field schema definitions, tags, and tag-content links.
Import MUST restore taxonomy before or together with content datasets so content references remain valid.

Dataset Note#

Library transfer SHOULD support attachment datasets (attachment metadata and content links).

Attachment Dataset Baseline#

Attachment metadata dataset file name: attachments.ndjson.
Each attachment row MUST include: id, content_id, uri.
Optional attachment row fields: name, mime, size, created_at.
content_id in attachments.ndjson MUST reference an existing content entry.

Validation and Safety Rules#

Import MUST reject package if any required dataset is missing.
Import MUST reject package if any dataset checksum validation fails.
Import MUST reject path traversal entries (../, absolute paths, or equivalent unsafe archive paths).
Import MUST validate schema and references before final commit.
Import MUST be atomic: on critical validation error, no partial final state is committed.

Import Order Baseline#

Validate package container and manifest integrity first.
Restore library metadata (library.json).
Restore taxonomy datasets before content datasets.
Restore content datasets.
Restore tags and tag-content links.
Restore attachments.ndjson after content datasets.
Run final referential validation and commit import atomically.

Settings Model

Mon, 01 Jan 0001 00:00:00 +0000

Settings Model#

Scope Groups#

account-scoped
library-scoped

Library-Scoped Settings#

default_status_on_create (enum)
default_priority_on_create (enum|null, default null)
allow_parent_links (boolean, default true)
archived_visible_by_default (boolean)

Account-Scoped Settings#

default_library_id (optional)
last_active_library_restore (boolean, default true)
auto_lock_timeout_minutes (integer, default 15, 0 means disabled)
require_unlock_on_start (boolean, default true)
sync_enabled (boolean, default false)
sync_mode (enum: manual|auto, default manual)
sync_interval_minutes (integer, default 15, used for sync_mode=auto)
export_compression_enabled (boolean, default true)
backup_enabled (boolean, default false)
backup_interval_hours (integer, default 24, used when backup is enabled)
backup_retention_count (integer, default 10)
timezone (text, default system timezone)
locale (text, default account locale or client/system fallback)

Fallback Rules#

If default_library_id is invalid or removed, fallback is last_active when available, otherwise explicit library selection.

TODO

Mon, 01 Jan 0001 00:00:00 +0000

TODO#

TODO: CLI commands;
TODO: CLI session lifecycle;
TODO: configuration and local storage;
TODO: errors and exit codes;
TODO: client-side security requirements.

Glossary

Mon, 01 Jan 0001 00:00:00 +0000

Glossary#

Fixed canonical terms used across the specification.

Account — top-level user domain context in core.
Session — authenticated runtime context bound to one account_id.
Library — structured storage domain for organized work.
Raw — account-scoped unstructured capture domain.
active_library_id — selected library for library-scoped operations.
library-scoped — operation/data requiring an active library context.
account-scoped — operation/data requiring only account session context.
unlock — session opening after credential validation.
sync — asynchronous data synchronization between local and remote runtimes.
promote — explicit transfer from Raw into Library.

Versioning Policy

Mon, 01 Jan 0001 00:00:00 +0000

Versioning Policy#

Scope#

This policy defines versioning rules for specification, core API, and project repositories.

Version Domains#

Spec Version - version of this specification.
Core API Version - version of runtime contract between clients and core.
Project Version - version of each concrete repository/package (core, cli, other clients).

Scheme#

All version domains MUST use Semantic Versioning: MAJOR.MINOR.PATCH.

SemVer Rules#

MAJOR: breaking contract change.
MINOR: backward-compatible feature addition.
PATCH: backward-compatible fix/editorial correction.

Compatibility Baseline#

Clients MUST declare supported Core API Version range.
Core runtime MUST expose current Core API Version.
On startup/handshake, client-core compatibility MUST be validated before normal operations.
Incompatible versions MUST fail with explicit compatibility error.

Release Mapping#

Spec Version MAY evolve independently of project package versions.
Breaking changes in spec core contract SHOULD increment both Spec Version and Core API Version major versions.
Project package versions SHOULD map to internal release cadence but MUST not bypass API compatibility checks.

RKL Format Versioning#

*.rkl format version is independent and MUST be declared in manifest.
Import MUST validate declared rkl version before processing datasets.

Changelog

Mon, 01 Jan 0001 00:00:00 +0000

Changelog#

Specification change history.

Rekeeply Specification on Rekeeply Spec

book

book#

article

article#

Content Typing

Content Typing#

Baseline Rules#

Default and Recommended Set#

book#

article#

Raw Model

Raw Model#

Purpose#

Contract#

Supported Payload Types#

Raw Item Storage Model#

Capture Inputs#

Metadata Enrichment#

Lifecycle and Operations#

Promote Flow Requirements#

Promote Mapping Rules#

Content Fields

Content Fields#

Field Classes#

Common Field Baseline#

system#

user_editable#

computed#

Attachment Model#

Rules#

Related Sections#

game

game#

Field Types

Field Types#

Scope#

Allowed Field Types#

screen

screen#

video

video#

quote

quote#

Tags

Tags#

Position in Model#

Tag Entity#

Content Association#

Editing Rules#

Import/Export and Sync#

Library Actions

Library Actions#

Scope#

Entry Actions#

person

person#

resource

resource#

Schema Actions

Schema Actions#

Scope#

Type Actions#

music

music#

audio

audio#

Presets

Presets#

Contract#

Operations#

Hierarchy Rules#

User Control#

Collections

Collections#

Contract#

Collection Fields#

Operations#

Hierarchy Rules#

misc

`book`#

`article`#

`system`#

`user_editable`#

`computed`#