TechHalla Profile picture
Sep 9 13 tweets 5 min read Read on X
X just open-sourced their algorithm code, and now you know how to position your content for max visibility.

10 keys to make your posts blow up, in this thread 🧵

1. Video is king, gets a 1.8x boost 👇
Include videos in your tweets.

Why? The algo auto-boosts them by 1.8x over text.

Use native X uploads, not YouTube links.

Hook in the first 3 secs with a question, bold claim or eye-catching content. Image
Image
2. Maximize Early Engagement

Aim for likes, RTs, and replies in the first 2 hours.

How?

Post when your audience is active (check analytics), notify loyal followers, and reply to early comments fast. Image
3. Build Coherent Communities (SimClusters):

Keep around 80% of your content focused on your main niche.

The algo assigns you to one of 145k communities based on who you follow and engage with.

Being consistent helps you show up in the right feeds and be seen as an authority. Image
Image
4. Leverage Graph Features

Create content that sparks interactions among connected users. "How many of your follows engaged?" is a huge signal.

More social proof = more visibility. Image
5. Strategic Timing Optimization

Post during your audience's peak hours. Algo boosts match time-of-day and day-of-week patterns.

Check X Analytics weekly and schedule top content there. Image
6. Generate Authentic Social Proof

Get to know your audience and learn from them: their comments, their interests, the kind of content they enjoy most...

Focus on that, keep driving engagement, and more and more people from their circles will start doing the same. Image
7. Avoid Safety Filters

Stay clean to dodge penalties. Algo's pToxicity/pNSFW models flag insults, edgy stuff.

Critique ideas, not people: "I disagree because..." Keep it family-friendly.

I keep AI debates constructive, no drama, steady growth.

One wrong word tanks reach. Image
8. Diversify Engagement Types

Don't just chase likes - mix replies, RTs, clicks. Different weights in the algo.

Add specific questions for replies, quotable insights for RTs, value-packed links for clicks. Image
9. Build Real Graph Score

Interact genuinely daily. Reply to 80% of comments, comment on niche peeps' posts.

Boosts future interaction probability, pushing your stuff higher in their feeds.

I spend 75 mins/day engaging AI community, pays off big.

Relationships > algorithms. Image
10. Optimize for Author Diversity

Don't flood - max 3-4 tweets/day, spaced 2-3 hours. Algo penalizes same-author overload.

Vary formats: video, poll, image, thread. Each post gets max individual reach.

Quality > quantity Image
These 10 steps are a summary of my detailed analysis based on the algorithm code that X just made public.

If you want to dive into all the details, I recommend this article I published recently (5–10 mins):
Hope this helps!

Follow me at @techhalla for more guides on X content creation, generative AI, and much more!

Drop a like, repost, and bookmark on the main post, it really helps me keep sharing this kind of content!

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with TechHalla

TechHalla Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @techhalla

Sep 11
I asked nano banana to handle my next marketing campaign and here’s what happened...

in just 5 minutes: branding, merch, and even an ad!

I’ll show you how to do it yourself in leonardo in this thread 🧵👇
First, I asked nano banana on @LeonardoAi_ to create a brochure with all the elements for my product.

From scratch. Here is the prompt:

Create a professional single logo, branding, packaging and merch brochure for a coffee brand called 'Viking beam'

I’m still blown away. Image
Image
Next, I took a photo of myself and the freshly generated brochure, and this time I asked nano banana to create a version of me enjoying the coffee.

Perfect! Prompt examples in the ALTs. the man on the photo is now at a top of a fjord bar terrace, wearing the merch of the brand. the product is over a wooden table next to a coffee maker. professional photography
the man on the photo is now at a bar, wearing the merch of the brand. the product is over a wooden table next to a coffee maker. professional photography
the man on the photo is now at his home, wearing the merch of the brand. the product is over a wooden table next to a coffee maker. professional photography
Read 6 tweets
Sep 9
save 100s of hours of image editing with Higgsfield's Banana Placement

all you need is one product shot!

full guide on how to create videos like this on higgs in minutes, in this thread 🧵👇
The team at @higgsfield_ai just dropped Banana Placement, a new feature in their Canvas that makes life much easier...

But first, let’s create our model.

initial image on the left, still pulled using nano banana on the right.

I’ve added the prompts in the ALTs. A candid low-angle iPhone shot of a stylish young woman sitting casually on a worn bench in an urban street setting, wearing oversized urban clothing and clearly visible Air Jordan sneakers. The dusk ambient lighting softly illuminates her natural skin texture, the fabric grain of her clothing, and casts gentle shadows, while street lights begin to glow in the background. In the foreground, a puddle on the asphalt reflects the scene with subtle distortions, reinforcing the gritty street style mood with cool tones and strong contrast. The spontaneous framing and slightly tilted angle capture...
macro close up of the sneaker. the person is standing
And now comes the magic.

We hit Edit, Banana Placement, upload the image we want to tweak into the Canvas, and paint a stroke over whatever we want to change.

Reflections included.

Upload the product (in this case, the new sneaker) and boom, you're done! Image
Image
Image
Read 8 tweets
Sep 6
concept rides inspired by iconic sneakers using nano banana.

examples and how to make your own in mins 🧵👇

1. Volkswagen Vans
2. Dodge Taylor
Citröen Suede
Read 10 tweets
Sep 6
if you don’t know how to use nano banana, you’re screwed in AI

8 tutorials to help you master the best image model out there 🧵👇🏻
1. Step inside your favorite video game.
Read 10 tweets
Aug 31
I asked nano banana to take me across Middle-earth and this is what happened...

all the details on how I made this video, below 🧵👇
Since I have nano Banana unlimited (I'll tell you how to get yours), I asked it to create Hobbiton for me in Google Maps mode.

Here’s the prompt:

dashcam google street view shot | Hobbiton streets | hobbits carrying out daily tasks like gardening and smoking pipes | sunny dayImage
I asked Grok to help me generate more images using the same prompt, but this time going through key moments from The Lord of the Rings.

It kinda got out of hand and I ended up making 38 stills in total.

Here are 4 of them, with the prompts in the ALTs.

More coming up. 👇 dashcam google street view shot | Seat of Seeing on Amon Hen | Ruined pavilion atop the hill, a hobbit-like figure from behind climbing the steps, the winding path down visible overlooking the river and lands beyond | panoramic view under emerging stars at dusk
dashcam google street view shot | High Pass in Misty Mountains | Gandalf leading travelers, dwarves mining nearby slopes | sunny day with mountain mist
dashcam google street view shot | Weathertop hill ruins | Rangers patrolling the area, ancient stones scattered around | clear sunny afternoon
dashcam google street view shot | Bree village streets | Aragorn watching from shadows, patrons entering the Prancing Pony Inn | sunny midday
Read 9 tweets
Aug 30
in this thread you’ll learn how to:

· create consistent characters
· come up with new angles for your scenes
· build impossible transitions

all with AI and in just minutes! Don’t miss it 👇
We're gonna use nano banana for the images and Kling for the animations.

All of it right on @freepik, where premium+ and pro plans get unlimited access to Google’s model!

Let’s start with the first image, the prompt’s in the ALT text. A photo of a Neanderthal man inside a prehistoric cave, dressed in rough fur clothing, his scarred arm extended as he touches ancient cave paintings depicting primitive hunting scenes | torchlight softly illuminates the ochre and charcoal pigments on the rock wall, casting warm shadows and revealing the texture of the stone | the cave interior is dark, quiet, and untouched by modern elements | captured with a Leica M11, 50mm lens, f/2.8, grainy film look, shallow depth of field
Starting from the image above, we’re gonna build out the rest of the story. Just upload the image and tell nano banana what changes you need!

Here’s an example for you. Image
Image
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(