Link: https://github.com/DisastrousClass/ao3-reading-companion

Built by Disastrous-Class9057 from reddit.

AO3 doesn't have a recommendation algorithm. This extension builds one for you. Completely privately, on your own device. With no information sent out.

Tell the extension what you love (tags, fandoms, authors). Browse AO3 normally. Over time, it learns what works for you and surfaces fics you'd actually enjoy reading, organized into a personal feed.

What you get:

• A multi-row feed with recommendations tailored to your taste. Including hidden gems, completed long reads, fresh chapters on your active reads, and discovery picks
• My Reading dashboard with Save for Later, currently Reading, and Read sections finally manage your AO3 reading queue properly
• Custom rows: define your own row like "Werewolf Long Reads, 50k+ words, complete, sorted by kudos" and the feed builds it for you
• "Why this?" transparency on every recommendation and see exactly which of your preferences matched
• Three-level dislike feedback (Never again / Less of it / Just from this author) that actually shapes future recommendations
• Manual update checks for fics you're actively reading
• Import your AO3 bookmarks to seed your Save for Later list
• Full preferences dashboard to fine-tune tags, fandoms, blocked content, structural preferences, and more

What it doesn't do:

This extension stores all data locally in your browser. Nothing is collected, transmitted, or shared. Not with Mozilla, not with AO3 beyond your normal browsing, not with the developer. Your reading habits and preferences are yours alone.

The extension respects AO3's servers with a 10-second spacing between requests and pauses if it ever hits a rate limit. It only operates when you're logged into AO3 account.

Requires Firefox. Works on any AO3 user account.

AO3 doesn't have a built-in recommendation algorithm, so fans have created several unofficial tools to help discover new fics. These range from browser extensions that learn your tastes locally to AI-powered apps and open-source projects.

Here’s a breakdown of the most notable unofficial AO3 recommendation engines and discovery tools.

🧩 Browser Extensions

These tools integrate directly into your browser and often work locally to protect your privacy.

• AO3 Reading Companion (Firefox): This extension builds a personalized recommendation feed based on your tags, fandoms, and authors. It runs entirely on your device, stores all data locally, and features a reading queue, chapter update tracking, and "Why this?" explanations for recommendations.
• AO3 Search Plus (Firefox): Unlike standard tag matching, this tool uses graph-based ranking to find relationships between works, tags, and authors. It helps surface related fics even when tagging is inconsistent or incomplete.
• AO3 Fic Recommender (Firefox & Chrome): This extension generates recommendations by analyzing a user's public bookmarks. It builds a local taste profile from fandoms, tags, and kudos to suggest works.
• AO3 Favorites Recommender (Greasyfork): A userscript that gives you a random work recommendation based on the tags you've marked as "favorite" on AO3. It's customizable and designed to be simple.

📱 Apps & Web Services

• AO3 Disco (Defunct): This was a popular, machine-learning-powered app that provided recommendations based on a fic you shared. It is no longer available, having shut down due to concerns from the AO3 team regarding privacy and third-party apps.
• FanFiction | AO3 unofficial: An iOS app that offers AI-powered story recommendations and conversational interactions about fics. It's a paid service ($44.99/year) and is an unofficial client for AO3.

🛠️ Technical & Open-Source Projects

For those who are technically inclined, several developers have created open-source scripts and tools, though many are experimental.

• Google Colab Notebook: A community-made Colab notebook designed to generate AO3 fanfic recommendations. It was created by a Reddit user as an alternative after AO3 Disco shut down.
• AO3 Graph Search (GitHub): An open-source browser extension that treats AO3 as a graph to discover works through query propagation rather than exact tag matching. It's a technical demo and explicitly states that no machine learning is used for its ranking.
• AO3 Recommendation Engine (GitHub): An older, content-based recommendation engine project from 2018. It scrapes AO3 data to form a database and generate recommendations.
• AO3 Recommender System (GitHub): A project built for a data science master's program. The goal was to build a set of recommendation tools, specifically for Marvel fanfiction on AO3.

⚠️ Important Considerations

When using these tools, it's helpful to keep a few things in mind:

• Privacy & ToS: AO3 is a non-profit archive that values user privacy and does not have an algorithm. Some tools, like the now-defunct AO3 Disco, have raised concerns about third-party access to user data. Always check a tool's privacy policy and data collection practices before installing it.
• Community Sentiment: Many users prefer AO3's manual, tag-based discovery system and are wary of algorithms, viewing the archive as a library rather than a social media feed.
• Experimental Nature: Many of the GitHub projects are proofs-of-concept or student projects. They may not be actively maintained or could have bugs.

If you'd like a deeper dive into any of these tools, feel free to ask.