Future of AnimeAPI

[nattadasu]

Note

Not-so-called-as-generative AI (Gemini 2.5 Pro) has been utilized to make this draft.

Hey everyone,

First off, a huge thank you to everyone who has used and supported AnimeAPI. It means a lot to me to see people finding it useful.

I'm writing this to let you all know that I've decided to rewrite the entire project from scratch. Honestly, the current code is really messy, which has made it incredibly difficult to maintain and fix bugs. For example, some of you may have noticed that certain titles like Everyday Host (MAL ID 60535) don't have related entries, even when the data exists from source (AOD, AniTrakt), and not to mention how long it would take for the script to make final result. Chasing down bugs like this in the current codebase is a nightmare.

Here's the plan for the new version:

1. A New Engine: The new backend will be built on top of my other project, Rensetsu, and especially its common library: librensetsu. This will provide a much cleaner foundation and allow me to use Rensetsu's database directly. My goal is to move away from AOD entirely to free the project from its restrictive licensing, making the API much friendlier for other developers to use.
   • As for context, Rensetsu tried to aim to be yet-another AOD-like database with better transparency and modularity. While I've wrote several site "modules", I'd admit it'd be tiring to do it alone.
   • And there's high chance that AnimeAPI will leveraging Rensetsu's scrapped data and extend more mappings compared to what I've done now.
2. Simpler Outputs: To deal with GitHub's repository size limits, the new API will only serve the final, processed data in JSON and TSV formats. Platform-specific files will be discontinued.
3. RAWs Remain: Don't worry, I will still provide the RAW data files from all sources. This is important for attribution or anything related to licensing, and that won't change.

Right now, I'm still figuring out the best way to roll this out—it might be a quiet, graceful upgrade, or it might be a completely new repository and domain. I'll keep you all updated on that front.

Don't worry, although the bug has been found on several titles, it wasn't enough to set the project to maintenance mode. However you can utilize other similar project like https://ids.moe by @tajoumaru (with reverse-compability with AnimeAPI) or https://github.com/BeeeQueue/arm-server temporarily.

This is a big undertaking for just one person, so I'm officially looking for help. If you're a developer who wants to see this rewrite happen swiftly and properly, I would be incredibly grateful for your contribution. Please feel free to reach out by commenting following discussion thread if you're interested in helping shape the future of AnimeAPI.

Thanks for your understanding and support,
Sultan Iskandar Maulana (@nattadasu)

[nattadasu]

Updates to the announcement

To ensure a stable transition and prevent breaking changes for existing users, we decided that the new version will be developed in a separate, temporary repository: AnimeAPI Next V4.

The current V3 API will remain operational and will source its data from the new V4 repository, retrofitting it to the V3 schema. This approach allows development to proceed on V4 while maintaining stability for V3.

---

Transition Plan & Infrastructure

• V3 Deprecation: The current V3 API is scheduled for deprecation around July 1, 2026.
• V4 Repository: Once V4 is stable, its code will be merged into a new branch in the current repository.
• API Versioning: A new X-ANIMEAPI-VERSION HTTP header will be implemented. A value of 4 will signify a request to the V4 schema.
• Performance: The V4 frontend server will be built using a compiled language supported by Vercel, such as Go, to improve response times. This will require learning the language to implement the change.

---

V4 Schema Preview

The following YAML provides a preview of the planned schema for V4.

• !optional indicates that a value/dict can be null.
• str? is an !optional string.
• int? is an !optional integer.

title: str
adb: int?			# renamed → .anidb
al: int?			# renamed → .anilist
ann: int?			# renamed → .animenewsnetwork
anisearch: int?
annict: int?
ap: !optional		# renamed → .animeplanet
  id: int	 		# NEW
  slug: str
dbti: int?			# renamed → .silveryasha
imdb: str?			# NEW
kaize: !optional
  id: int?			
  slug: str  		# previously simply named as .kaize
kitsu: !optional
  id: int
  slug: str			# NEW
lc: int?			# renamed → .livechart
mal: int?			# renamed → .myanimelist
nautiljon: !optional
  id: int
  slug: str	    	# previously simply named as .nautiljon
notify: str?
otakotaku: int?
shikimori: str?		# NEW, taken directly from Shikimori
simkl: int?
syoboi: int?		# renamed → .shoboi
tmdb: !optional		# renamed → .themoviedb
  id: int
  type: str			# tv | movie
  season: !optional	# only provided if type == tv
    id: int			# NEW, season ID
    number: int		# NEW, reflects Trakt's season number
trakt: !optional
  id: int
  slug: str
  type: str			# shows | movies
  season: !optional	# only provided if type == tv
    id: int			# NEW
    number: int
tvdb: !optional		# NEW, only for tv shows
  id: int
  season_id: int?	# Depends on TMDB/Trakt has the data

---

Summary of Changes

What's being added or changed

• New Schema: A more structured and expanded data schema (detailed above).
• Expanded Platform Support: Mappings for IMDb, TMDB, and TVDB will be included, sourced from extended AniTrakt datadump.
• Standardized Keys: Common site names are now abbreviated for consistency (e.g., myanimelist is now mal).
• Licensing: The source code will adopt a more permissive license, and the data output will be licensed under ODbL 1.0.

What's being removed

• Platform-Specific JSONs: V4 will no longer provide individual JSON files. Instead, it will serve a raw SQLite database and a TSV file.

---

Important caveats

1. Please be aware that while TVDB mappings are provided, we cannot fetch data directly from their API due to its paywall. All TVDB information is sourced exclusively through Trakt.
2. On Kaize, the ID for some niché/less popular titles can be nulled as to get the ID, it requires to iterate anime detail page one-by-one, which definitely will exhaust server time.