How the X algo Ranks Your Posts

This is the For You feed stack, split into 4 big parts:

→ home-mixer = It pulls everything together and decides what shows up on for you

→ thunder = This handles posts from people you already follow. Familiar, low-risk content.

→ phoenix = This finds posts from people you don’t follow and figures out which ones you’re most likely to care about using Grok.

→ candidate-pipeline = This is the assembly line. It:

The README just walks through that entire process step by step from opening the app to the posts you actually see.

Quick Detour for my sponsors today: DEEL

Check out what they are offering for my readers:

Deel's free 2026 trends report cuts through all the hype and lays out what HR teams can really expect in 2026. You’ll learn about the shifts happening now, the skill gaps you can't ignore, and resilience strategies that aren't just buzzwords. Plus you’ll get a practical toolkit that helps you implement it all without another costly and time-consuming transformation project.

Get the free report today.

When someone opens the For You, this is how the algo evaluates available posts:

1) Query hydration

They fetch your:
→ recent engagement history
→ following list + preferences

2) Candidate sourcing (get posts to consider)

Two streams:
→ Thunder (in-network)
Posts from accounts you follow, served fast from an in-memory store.

→ Phoenix Retrieval (out-of-network)
It does two things at the same time:

Then it compares the two and asks:
“Which posts look the most like things this person usually engages with?”

The closer the match, the higher those posts get pushed into your feed.

3) Hydration (enrich candidates)

Adds metadata:
→ text/media
→ author info
→ video duration
→ subscription eligibility

4) Pre-scoring filters (remove junk before spending ML)

Filters include:
→ duplicates
→ too old
→ your own posts
→ blocked/muted authors
→ muted keywords
→ previously seen / previously served
→ subscription ineligible content

‼️ 5) Scoring (the “ranking brain”)

This is the part that actually decides what gets shown and what disappears.

Grok looks at a post and asks:
“What are the chances someone will actually do something with this?”

Not just "likes", it predicts:

→ will they stop and read
→ will they reply
→ will they share it
→ will they click the profile
→ or will they mute, block, or hit ‘not interested’

Every possible reaction gets weighed. Posts that are likely to get good reactions move up. Posts that are likely to get bad reactions get pushed down or removed.

So your reach isn’t random.

it’s the system constantly guessing how people will react to you before your post even finishes circulating.

And then they layer on:

6) Select top K + post-selection filters

Final cleanup includes:
→ visibility / safety filtering
→ conversation deduping (don’t show 7 branches of the same argument)

Positive
→ favorite/like, reply, repost, quote
→ click, profile_click
→ video_view, photo_expand
→ share, dwell
→ follow_author

Negative
→ not_interested, block_author, mute_author, report

So yes: the system is literally designed to reward what people are likely to enjoy and bury what they’re likely to reject. (are we surprised?)

If you want your posts to show up on for you, optimize for:

1) Dwell + completion
→ first 2 lines must create a pause
→ scannable formatting keeps them reading
→ video needs retention, not just a view

2) High-intent actions
Likes are cool. But these are the real engagement signals you should optimize your content for:

→ replies that aren’t one-word junk
→ profile clicks
→ follows
→ shares / DMs

3) Avoid negative reactions
If your content causes “not interested”, mutes, blocks you'll receive "negative scoring".

4) Stop spamming your own reach
The repo explicitly includes an Author Diversity Scorer. Translation: you showing up repeatedly (posting toooo much) gets devalued. Nobody needs to post more than 8-10 times a day IMO.

“When I first opened Attio, I instantly got the feeling this was the next generation of CRM.”
— Margaret Shen, Head of GTM at Modal

Attio is the AI-native CRM for modern teams. With automatic enrichment, call intelligence, AI agents, flexible workflows and more, Attio works for any business and only takes minutes to set up.

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉 Start for free today

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

Reply and ask!

I fed the entire X algo code to my AI and I can ask it anything you need.

🔖 BTW for more breakdowns like this, make sure you’re following me on 𝕏: @aunysillyme

You can also hop on a call with me if you need help with your content strategy, monetization or anything related to social media and X in general:

SEE YOU ON THE NEXT ONE
- AUNY 🧡