The Tafo Engine uses advanced AI techniques to recommend highly relevant articles from your own archive and helps drive up engagement.

tafo logo
Set me up!

Get Started Easily!

1. Embed this between the <HEAD> tags in your article template

<script src="https://s3.amazonaws.com/tafo.ai/yoursiteid/tafo.js"></script>

2. Paste this in the <BODY> of the article template where you want related content to show up

<div id="tafo_related"><!-- this will be replaced dynamically by TAFO --></div>

3. Contact us to whitelist your domain and complete your setup!

❤️ Crafted in San Francisco!

Index your archive, automagically!

The Tafo Engine implicitly learns your readers' preferences to better predict the kind of content that is going to engage, challenge and inform them. Editors can choose to bias the content towards specific topics of interest to your audience.

Readers spend more time engaging with your quality content, increasing loyalty and brand perception.

Customized to reader & editor preferences

Increase your engagement & loyalty

Articles from your own archive are automatically indexed and displayed by contextual meaning.  This means an article about “meditation” gets recommended when reading one on “mindfulness”, “vipassana”, “zen”, etc.





Monthly Page Views


1 M

Monthly Unique Visitors


Indexed Articles









5 M

10 M




25 M

50 M

200 M

50 M

3 M




Extra 5M Page Views

Extra 50k Articles Indexed



One-Time Setup





How it works

Crawling & Vectorization

The Tafo Engine periodically scans your publication for new articles and automatically vectorizes it along hundreds of dimensions that are implicitly discovered and maintained for your archive.

Indexing and Real-time Recommendations

The entire vectorized index of your archive is loaded and updated in memory  to serve real-time recommendations using different signals from the context - article being viewed, readers' history, editorial biases etc.

Automatic Feedback Loop

The engine tracks how the recommendations perform, automatically improving itself over time, and tailoring itself to your specific content and readers.

Fix the following errors: