philz1337x
@philz1337x
Followers
11,337
Following
184
Media
177
Statuses
984
AI developer building open source AI apps 💎 $5K/m 🕵️♂️ free 🔔 free
Thailand
Joined January 2023
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
TLP LETS GET BACK TOGETHER
• 803890 Tweets
ارسنال
• 719764 Tweets
Madison Square Garden
• 450817 Tweets
Arsenal
• 312264 Tweets
Liverpool
• 180241 Tweets
Saka
• 114774 Tweets
Eagles
• 92919 Tweets
Gabriel
• 72626 Tweets
Lions
• 63644 Tweets
#ARSLIV
• 60046 Tweets
Tarcísio
• 39031 Tweets
Bengals
• 36806 Tweets
Merino
• 34637 Tweets
#deprem
• 30285 Tweets
Arteta
• 29654 Tweets
Titans
• 28742 Tweets
Trent
• 26056 Tweets
Falcons
• 24601 Tweets
Havertz
• 22033 Tweets
Jalen
• 21290 Tweets
#InterJuve
• 21186 Tweets
Robertson
• 21089 Tweets
Martinelli
• 21061 Tweets
Colts
• 20297 Tweets
Kerem Aktürkoğlu
• 17489 Tweets
Danilo
• 15339 Tweets
Jameis
• 14880 Tweets
Betis
• 13531 Tweets
Konate
• 12386 Tweets
Kiwior
• 11219 Tweets
Pinned Tweet
🥳 I transformed from a corporate job I didn’t like to building my own app, , making $5,000 MRR.
🛠️ I worked as a mechanical engineer in Switzerland and made $100k per year, but it was so boring. Everything was so slow, and my tasks felt useless because I
✨ Open sourcing my AI image upscaler helped it reach $2,800 MRR
I worked on it for the last 6 months but it did not really take off, so I was about to give up
Then last month, I decided to just open source it and see what would happen, it took off and
68
49
704
97
40
1K
✨ For the last few months I have been reverse engineering Magnific AI's famous upscaler
It uses MultiDiffusion, ControlNet tiles and details LoRas
In true AI spirit, I am open sourcing it for everyone to use for free in your apps
It's designed to easily control the amount of
166
376
3K
✨ I upscaled a 64x64 pixel image to 37 Megapixels 🤯
Workflow: the new Clarity AI Low Res for upscaling from 64 to 1024px and Clarity AI regular for 1024 to 6144px.
I made both models open-source and am now implementing the low-resolution option to my app
47
157
1K
I had a great time working for
@levelsio
on his AI models for Photo AI.
He inspired me to realize that I don’t have to follow the typical corporate life, but can continue doing what I loved in my childhood—coding and working independently on my laptop—and actually achieve
22
26
941
I make $5k per month with an app that's fully open-source.
Why do people pay for something that's free to use?
Ready-washed and chopped salad is a good example. It can cost 4 times as much as raw salad, but people like to pay for it because it's worth it to save time and
🥳 I transformed from a corporate job I didn’t like to building my own app, , making $5,000 MRR.
🛠️ I worked as a mechanical engineer in Switzerland and made $100k per year, but it was so boring. Everything was so slow, and my tasks felt useless because I
97
40
1K
22
31
748
✨ Open sourcing my AI image upscaler helped it reach $2,800 MRR
I worked on it for the last 6 months but it did not really take off, so I was about to give up
Then last month, I decided to just open source it and see what would happen, it took off and
68
49
704
✨ Free ComfyUI Workflow for Clarity Upscaler!
Finally, we have a fully ComfyUI workflow that runs locally on your own GPU and uses the same technology as Clarity
To use it, just download the free-workflow.json file from the ComfyUI-ClarityAI repo
😀 Ad: If you're looking for
✨ Clarity AI upscaler is now available in ComfyUI! 🤩
It was the most requested feature since I launched Clarity AI.
At first, I tried to reproduce my whole workflow in ComfyUI, but it seemed impossible. However, now the node uses the Clarity AI API, so users have all the
21
69
469
22
100
684
Mona Lisa Reimagined in Different Ethnicities.
✨Stable Diffusion + ControlNet
More images in the thread 👇
21
56
447
Get $100 if I skip a workout 💸
I’m done with excuses. No more “I’m not in the mood” or “I’m in a new place.” From now on, it’s 4 workouts a week, no exceptions.
Here’s the deal:
For every workout I skip in the next 6 months, I’ll pay $100 to a random person who engages with
358
7
497
✨ Clarity AI upscaler is now available in ComfyUI! 🤩
It was the most requested feature since I launched Clarity AI.
At first, I tried to reproduce my whole workflow in ComfyUI, but it seemed impossible. However, now the node uses the Clarity AI API, so users have all the
✨ I upscaled a 64x64 pixel image to 37 Megapixels 🤯
Workflow: the new Clarity AI Low Res for upscaling from 64 to 1024px and Clarity AI regular for 1024 to 6144px.
I made both models open-source and am now implementing the low-resolution option to my app
47
157
1K
21
69
469
✨ I increased the max resolution on my upscaler app to 13,000 x 13,000 pixels (or 176 megapixels) 🤯
This makes it the highest resolution AI upscaler available today
It takes 19 minutes to run on an A100 GPU and produces a 270MB PNG file 😱
I think
32
46
481
✨ Open-source Video-Style-Transfer model.
I have built a replicate model around Diffutoon from
@zhouwenmeng
and his team, which is a toon shading model built on ControlNet lineart and SD 1.5.
You can use it locally for free by cloning the GitHub code, on Replicate as an API,
11
84
413
🚅 Upscaling images is 50% faster now 🤯
Only takes 12s to upscale from 1 -> 4 Megapixel 😱
How? I just implemented xformers library 😂
App:
Code:
API:
17
67
410
🧑💻 How to Get a Job as an AI Image Developer or Build Your Own AI Image Apps
I quit my $100k engineering job to make it as a generative AI developer. There are many ways, but here is how I have been doing it for the last 12 months:
1. Create AI images on your computer with
17
40
356
🔍 Multi-step image upscaling 😱
From 1 to 38 Megapixels in a single click 🤯
Until now, only x2 upscales gave high-quality results, so I implemented multi-step upscaling, which splits high multipliers like x16 into multiple x2 upscales internally
Now live:
😀User-friendly
20
50
347
🧑🎨 I'm developing an open-source Style-Transfer!
Apply the style from one image to another.
It uses ControlNet Canny to preserve structure and IP-Adapter to apply the styling.
Still needs some work for certain images, but you can start using it now:
✨ As an app:
19
41
325
✨ Clarity AI can now upscale to 205 megapixels (14,336 x 14,336 pixels) 🤩
The file size of the .PNG is about 400 MB, and the .JPG is even 90 MB 😲
This higher upscale resolution is now possible because I removed a pixel size limitation in the PIL image library, which thought
✨ I increased the max resolution on my upscaler app to 13,000 x 13,000 pixels (or 176 megapixels) 🤯
This makes it the highest resolution AI upscaler available today
It takes 19 minutes to run on an A100 GPU and produces a 270MB PNG file 😱
I think
32
46
481
14
34
331
✨ControlNet and Stable Diffusion in restoring a stained blurry 70 year old photo
5
16
290
🤩 75% cheaper Image upscaling 🤑
GPU costs for image upscaling from 1 to 4 Megapixels are less then a penny now 💵
Since the speed improvements I made, Clarity Upscaler now runs just as fast on both the cheaper A40 and the pricier A100 GPUs. The total GPU costs have decreased
20
28
315
🌺Enhanced detail with Fractality in Clarity-Upscaler 💎
Fractality controls the level of detail in an image. Lower values provide a clear, broad perspective, while higher values uncover rich, intricate details.
For complex scenes, a higher Fractality enhances the depth, but
16
42
268
Clarity AI is open source and you can host it yourself and run it for free, or run it on Replicate
If you want to use it without writing any code, I have my own hosted and paid version at which helps support my work as an AI dev
It's 25% cheaper than
✨ For the last few months I have been reverse engineering Magnific AI's famous upscaler
It uses MultiDiffusion, ControlNet tiles and details LoRas
In true AI spirit, I am open sourcing it for everyone to use for free in your apps
It's designed to easily control the amount of
166
376
3K
7
21
260
Free credits lower my conversion rate by 60%.
I conducted an A/B test to check if giving free credits could increase my revenue.
These are the results:
- Conversion rate dropped from 2.1% to 0.9% (-60%) 🛑
- Revenue after free credit costs per user dropped from $0.38 to $0.23
35
10
255
🤩 30% more speed and increased quality in high-resolution image upscaling
I've improved clarity upscaling model by reducing the creativity by 20% in every internal upscaling step.
It makes the amount of hallucination much more controllable, thus increasing the quality while
23
23
236
💸 Free Anime Image Upscaling 🙀
Now you can upscale anime images with the open-source Clarity Upscaler!
How to use it:
😃Easiest:
Choose the Anime preset on
⚙️API for integration in your app:
🛠️ Cog:
13
30
231
🔓 Open-sourcing gave it a lot of social proof, which I have now added to the website
- 1.8 Million runs on
@replicate
- Featured in apps like from
@levelsio
- Many Testimonials like
@rrmdp
for or
@JDL4_
13
16
211
✨ Launching another open-source app:
It creates wallpapers for 📱 smartphones, 💻 laptops, and big wide screens up to 5K.
Using the Stable Diffusion model SDXL RealisticVision V3, it creates a base image and upscales it with my Clarity Upscaler.
I
16
16
194
🤑 I've reduced pricing for Clarity AI!
It's now ONLY $19/mo, or 50% cheaper than Magnific AI, with same amount of upscales
I realized public models are cheaper to run on my provider, Replicate, so after fully open-sourcing Clarity, I was able to cut costs
✨ Use it as an app
✨ For the last few months I have been reverse engineering Magnific AI's famous upscaler
It uses MultiDiffusion, ControlNet tiles and details LoRas
In true AI spirit, I am open sourcing it for everyone to use for free in your apps
It's designed to easily control the amount of
166
376
3K
12
11
161
a1111 webui Parameters:
Prompt:
masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>
Negative prompt: (worst quality, low quality, normal quality:2) JuggernautNegative-neg
Steps: 18,
Sampler: DPM++ 3M SDE Karras,
CFG scale: 6.0,
Seed: 1337,
13
7
133
🎨 LoRas now fully available in Clarity Upscaler!
Use any LoRa from Civitai or Hugging Face with your upscale, like this Chinese ink style, for example.
Live open source on GitHub, Replicate, and in the ClarityAI app.
6
17
130
🔍 I’ve added optional pre-downscaling for Clarity AI
It increases quality and speed for images with high resolution but low quality like old game assets
🛠️ Open-Source:
⚙️ Replicate API:
🤑 I've reduced pricing for Clarity AI!
It's now ONLY $19/mo, or 50% cheaper than Magnific AI, with same amount of upscales
I realized public models are cheaper to run on my provider, Replicate, so after fully open-sourcing Clarity, I was able to cut costs
✨ Use it as an app
12
11
161
6
8
133
🥳 The Clarity-upscaler model crossed 2 million runs on
@replicate
After less than 3 months, it's now in the "Popular models" list alongside big companies like Meta, Tencent, Stability-AI, and ByteDance.
✨ For the last few months I have been reverse engineering Magnific AI's famous upscaler
It uses MultiDiffusion, ControlNet tiles and details LoRas
In true AI spirit, I am open sourcing it for everyone to use for free in your apps
It's designed to easily control the amount of
166
376
3K
11
3
118
🐦 Transform images with any Stable Diffusion model 🎨
Now you can upscale with any safetensors file from platforms like Huggingface, Civitai, and more
In this example, I used "Disney Pixar Cartoon Type A" from Civitai
App:
API:
2
14
117
✨ The AI image upscaler now has its own API ⚙️
It's great for business customers because:
- The API can respond with the finished image, so there's no need to handle webhooks
- Output files are permanent
- Simple parameters: set creativity, resemblance,
🔍 Multi-step image upscaling 😱
From 1 to 38 Megapixels in a single click 🤯
Until now, only x2 upscales gave high-quality results, so I implemented multi-step upscaling, which splits high multipliers like x16 into multiple x2 upscales internally
Now live:
😀User-friendly
20
50
347
8
15
111
✨ Create high-quality images now with Clarity AI
It uses Flux schnell from
@bfl_ml
to create a base image and upscales it with Clarity.
The base image only takes 1 second 🤯 and has great prompt adherence. The upscaling step adds details and makes it high quality.
Prompt:
3
12
104
How to Upscale Images for Free with Clarity Upscaler:
1. Install A1111 WebUI (three options from hardest to easiest):
a) Follow the official tutorial:
b) Watch the YouTube tutorial:
c) Use One-Click start on
@runpod_io
:
I make $5k per month with an app that's fully open-source.
Why do people pay for something that's free to use?
Ready-washed and chopped salad is a good example. It can cost 4 times as much as raw salad, but people like to pay for it because it's worth it to save time and
22
31
748
5
2
104
✨ Add sound to your videos with AI
I am launching Video2Sound and sharing it as open-source.
At its core, I am using FoleyCrafter from
@zengyh1900
and her team. I have made some adjustments and built a
@replicate
model around it.
Everyone can use it with a very simple UI on
✨ I added a sharp effect to , which makes the upscaled image a bit crisper.
👩💻 App:
⚙️ API:
🛠️ Open-Source Code:
🔧 Replicate:
3
11
77
9
10
97
Most upscalers use a checkpoint trained on general images, that's why they have a high resemblance loss.
But with Clarity, you could use a dreambooth model or a LoRa and upscale with that and should reach very high resemblance.
The biggest downsides of all open-source upscalers are that they have a HUGE resemblance loss, making them unusable for any real human/Dreambooth-related stuff (LEFT).
We've internally built an upscaler that has <5% resemblance loss (RIGHT)
9
5
107
4
5
83
✨ Increased monthly revenue by +60% by adding a one-time payment option to
A highly requested feature was the ability to upscale just one image without a subscription. So, I added Clarity AI Go, which is a $5 one-time payment option to upscale one image
11
3
84
🚄 Speed up your Stable Diffusion code 💨
There are many ways to speed up SD in A1111 and diffusers code. It depends on the use case, hardware, etc., which options work. Here is a list of the options I know.
If you feel overwhelmed by the number of options, try only the first
🚅 Upscaling images is 50% faster now 🤯
Only takes 12s to upscale from 1 -> 4 Megapixel 😱
How? I just implemented xformers library 😂
App:
Code:
API:
17
67
410
2
15
84
🤩 This could also be a great feature for
I am starting now to try to add this
I'll pay you $10,000
If you can make a Replicate model (not ComfyUI, not anything else, ONLY a Replicate model)
That takes an image
Detects the hands area
Fixes the hands consistently:
- regular hands
- crossed hands
- hands facing camera
And is fast like 10-20 seconds
183
46
930
6
4
78
✨ I added a sharp effect to , which makes the upscaled image a bit crisper.
👩💻 App:
⚙️ API:
🛠️ Open-Source Code:
🔧 Replicate:
✨ I added Video Style-Transfer to
You can convert an existing video into another style, like anime.
The hard part was achieving consistency and no flickering in the video.
At the moment, it takes 30 minutes to process a 30-second video. 😅
Now I am
4
3
51
3
11
77
💸 Users now get more image upscales on !
Thanks to cost reduction, I've increased the amount of image upscales by 50% for Premium and 100% for Business plans
Clarity is now up to 74% cheaper than competitors
🙂In addition, I added yearly options with 2
🤩 75% cheaper Image upscaling 🤑
GPU costs for image upscaling from 1 to 4 Megapixels are less then a penny now 💵
Since the speed improvements I made, Clarity Upscaler now runs just as fast on both the cheaper A40 and the pricier A100 GPUs. The total GPU costs have decreased
20
28
315
2
4
76
🎨 Switched to the open-source style transfer of
@fofrAI
Apply the style from one image to another
@fofrAI
's version uses SDXL with ControlNet depth and IP-Adapter. It's much faster than my style transfer model based on SD1.5 and is always warm on Replicate 🔥
⚙️ Replicate:
I've pushed a new style-transfer model to Replicate. It's a wrap around the fantastic IPAdapter work
@cubiq
has been doing
All images use my pfp as a style input with a simple text prompt (living room, sports car, bird, coffee)
Links to model, repo and workflow below 👇
6
12
137
5
6
71
✨ Should I add relighting an image with IC-Light to ?
IC-Light (Imposing Consistent Light) is a tool to manipulate the illumination of images.
As far as I understand, it generates a new background, changes the lighting of the foreground, and combines
✨ Clarity AI upscaler is now available in ComfyUI! 🤩
It was the most requested feature since I launched Clarity AI.
At first, I tried to reproduce my whole workflow in ComfyUI, but it seemed impossible. However, now the node uses the Clarity AI API, so users have all the
21
69
469
4
4
60
🤩 Resemblance is solved on Clarity Upscaler
@levelsio
is using dreambooth files from to keep resemblance
Clarity supports dreambooth and LoRa files to archive high resemblance
I think this should also work with product photos 👀
✨ running now with
@philz1337x
's Clarity AI upscaler
I'm now integrating it into Photo AI, it takes some work because I have to convert the Dreambooth files (of trained people) from .tar to .safetensor and then add Clarity AI at the end of the processing
30
7
267
2
1
58
I got 10k followers in 6 months with basically only 5 posts.
They fall into two types:
- Sharing something valuable for free
- Celebrating a desirable success
I will put the 5 posts in the comments 👇
You only need to write 26 banger tweets
And you'll have 100,000 followers
116
21
1K
5
5
56
✨ I added Video Style-Transfer to
You can convert an existing video into another style, like anime.
The hard part was achieving consistency and no flickering in the video.
At the moment, it takes 30 minutes to process a 30-second video. 😅
Now I am
4
3
51
4
1
54
✨ I added Video Style-Transfer to
You can convert an existing video into another style, like anime.
The hard part was achieving consistency and no flickering in the video.
At the moment, it takes 30 minutes to process a 30-second video. 😅
Now I am
✨ Clarity now supports pattern upscaling 🔍
The problem with patterns is that the edges should fit perfectly with the other opposite edges. When you upscale a pattern usually the image changes a bit and it stops being a seamless pattern.
Thanks to
@pixelprotest_
's work Clarity
7
3
34
4
3
51
I was talked into doing deadlifting by
@levelsio
, but I preferred running and yoga.
But now that I broke my toe, running is impossible.
I realized lifting weights is fine with the toe, and it gives me the same effect of clearing my mind.
6
1
47
I got 3 million views in three months 🤯
Most of the views came from two types of posts:
👉 Giving something away for free, like open-sourcing my image upscaler, knowledge, or a "how to become an AI developer" tutorial.
👉 Sharing a big success, like when I reached $5,000 MRR.
9
0
44
Version is live!
Code:
Replicate:
3
4
41
🤯 Clarity-Upscaler crossed 3.6 million runs on
@replicate
.
It's now one of the three most used upscalers, together with Real-ESRGAN and SwinIR.
🥳 The Clarity-upscaler model crossed 2 million runs on
@replicate
After less than 3 months, it's now in the "Popular models" list alongside big companies like Meta, Tencent, Stability-AI, and ByteDance.
11
3
118
5
3
40
FLUX has beaten Stability AI in just 3 days 🤯
That's how long it took the newest AI image model, FLUX, to overtake the runs of the 3-month-old Stability AI model, SD3 🥳
Congrats on the huge success of
@bfl_ml
(Black Forest Labs), which consists of the legendary Stable
✨ Create high-quality images now with Clarity AI
It uses Flux schnell from
@bfl_ml
to create a base image and upscales it with Clarity.
The base image only takes 1 second 🤯 and has great prompt adherence. The upscaling step adds details and makes it high quality.
Prompt:
3
12
104
2
0
37
Here wa go dghin. 😉
Saying that you have reverse-engineered Magnific and then showing something using a completely different pipeline, and therefore with worse results, is like saying that I have reverse-engineered Sora using Animate Diff and a chili pepper 😂
25
6
156
4
0
36
✨ Clarity now supports pattern upscaling 🔍
The problem with patterns is that the edges should fit perfectly with the other opposite edges. When you upscale a pattern usually the image changes a bit and it stops being a seamless pattern.
Thanks to
@pixelprotest_
's work Clarity
You can now earn $$$ with 🤑
I have added an affiliate program, so you get 30% of all sales for the first three months of every referred customer. That's up to $490 for one customer with a business plan and $26 for a pro plan.
I think it’s a great
1
2
20
7
3
34
🤩 Outputting a .jpg instead of a .png speeds up prediction time on GPU
✨Added an option to choose the output file format from [jpg, png, webp] on my image upscaler, Clarity AI
I always thought it's better to do image conversion in the backend because CPUs are cheaper than
✨ The AI image upscaler now has its own API ⚙️
It's great for business customers because:
- The API can respond with the finished image, so there's no need to handle webhooks
- Output files are permanent
- Simple parameters: set creativity, resemblance,
8
15
111
2
1
34
✨ I built a free notification app called 🔔
If something 🚨 breaks in my apps, I need to know instantly so I can fix it as soon as possible, and I also want to stay informed about what's happening
To address this, I built SendAlert, which allows me to
7
3
34
@levelsio
I implemented this for Clarity:
1. It takes the hand area.
2. Upscales and enhances it.
3. Stitches it back into the original image.
It does enhance the hands, gives them more details, and fixes them a bit.
However, they are not fully fixed and still have anomalies.
I have
3
3
33
@levelsio
Thanks. One thing that let me work pragmatically is that I didn’t learn it in university or at companies.
1
1
33
@NathanSRobinson
The base tech is open source and it just needed to put puzzle pieces together
1
0
32
Clarity Upscaler is now 40% faster 🏎️
Just switched GPU hardware on Replicate from A40 to A100.
Downside is a 20% cost increase, but I'd say it's worth it.
2
1
32
Creative upscaling works so well for illustrations
✨ Clarity AI can now upscale to 205 megapixels (14,336 x 14,336 pixels) 🤩
The file size of the .PNG is about 400 MB, and the .JPG is even 90 MB 😲
This higher upscale resolution is now possible because I removed a pixel size limitation in the PIL image library, which thought
14
34
331
2
2
34
I am not afraid of copycats.
You could build a great app with the best and newest tech, but there will still be copycats. When people don't know you and see a copycat first, they will pay without comparing.
So first, it's about visibility and getting users, and open-sourcing
I make $5k per month with an app that's fully open-source.
Why do people pay for something that's free to use?
Ready-washed and chopped salad is a good example. It can cost 4 times as much as raw salad, but people like to pay for it because it's worth it to save time and
22
31
748
1
1
32
Every developer trying to make 💵:
100% coding, 0% marketing = $0 MRR
90% coding, 10% marketing = $500 MRR
Mmmh... got a feeling...
But first, let me code this delete button feature 👨💻
🗑️ Added a [Delete] button
It was the feature most users asked for 😅
Now you can delete any upscale or style-transfer you want to remove from your feed
0
0
22
4
4
29
⚠️ The "Iron Triangle" is a myth created by slow project managers and corporate workers.
It claims you can't have quality, speed, and cost-effectiveness all at once, and not even maximize two of them.
This is outdated because:
- Tools like ChatGPT have greatly increased 📈
5
4
28
Works! It doesn't fully fix the hands, but it improves them a lot, I think.
7
0
28
Why taking cold showers can help grow your business. See below 👇
6
2
22
Imagine your product is so good, people buy it even when they struggle to open it.
8
0
26
🥳 New open-source image model Flux claims to outperform Midjourney 🤩
From legend
@robrombach
and his team.
🛠️ Code:
⚙️ API:
We are excited to announce the launch of Black Forest Labs. Our mission is to develop and advance state-of-the-art generative deep learning models for media and to push the boundaries of creativity, efficiency and diversity.
132
298
2K
2
3
25
Style transfer with Flux 🤩
FLUX is incredible at prompt following, but can't do Style Transfer yet
here's a neat trick: you can use FLUX Schnell's gens as Controlnet into an SDXL workflow with an IPA + Style Ref, boom, you now both style and prompt following :)
some examples + link to a glif workflow:
8
37
301
0
2
24
🤯 Studies show that we become happy in nature because of a specific bacterium! 🍀
Mycobacterium vaccae 🦠, also known as "Psychobacter vaccae," is found in soil and dirt. Research suggests this bacterium can boost our health 🌱, particularly by improving our immune system and
3
3
23
🛑 My app failed.
I'm closing because it doesn't really make money.
I think if people want to create a wallpaper, they just do it themselves with an AI image generation tool.
Since I have implemented image creation into Clarity AI, it's also not
✨ Launching another open-source app:
It creates wallpapers for 📱 smartphones, 💻 laptops, and big wide screens up to 5K.
Using the Stable Diffusion model SDXL RealisticVision V3, it creates a base image and upscales it with my Clarity Upscaler.
I
16
16
194
8
1
23
🐌 Uploading a 1-megapixel image takes >20s??
File upload with Firebase Storage and Google Cloud is now slower than the actual upscaling process.
Any ideas on how I can fix it for a better user experience?
5
0
24
✨ I spent almost my entire first MRR on upgrading my domain to
It was .cc before, and .co seems to be more common, more similar to .com, and easier to recognize
✨ Open sourcing my AI image upscaler helped it reach $2,800 MRR
I worked on it for the last 6 months but it did not really take off, so I was about to give up
Then last month, I decided to just open source it and see what would happen, it took off and
68
49
704
1
0
22
🗑️ Added a [Delete] button
It was the feature most users asked for 😅
Now you can delete any upscale or style-transfer you want to remove from your feed
0
0
22
I switched to
@Cloudflare
R2 and it takes 84% less time to upload! 😯
Google storage: 25.6s
R2: 4.1s
Only 4s now, maybe enough?
2
1
22
You can now earn $$$ with 🤑
I have added an affiliate program, so you get 30% of all sales for the first three months of every referred customer. That's up to $490 for one customer with a business plan and $26 for a pro plan.
I think it’s a great
🥳 The Clarity-upscaler model crossed 2 million runs on
@replicate
After less than 3 months, it's now in the "Popular models" list alongside big companies like Meta, Tencent, Stability-AI, and ByteDance.
11
3
118
1
2
20
Me in the past: 👨💻
Coding for a week and only telling another nerd what I did
Me now: 🥳
Choosing another GPU on Replicate and making a huge announcement with video and 10 emojis 😄
1
0
20
I tried to switch to Stable Diffusion 3 for Wallpaper AI, but the results are not good 😳
SD3 👈 👉 RealisticVisionXL
✨ Added an option to generate unlimited wallpapers for one prompt on
The option [fixed] uses 1337 as the seed, and [random] creates a random one
So if you find a good prompt, you can switch to random and create similar wallpapers until you find your 👍
0
2
15
7
2
20
For a very user friendly Interface but paid:
2
1
20
🔍 Increased the quality with
1
0
20
The GitHub repo is an AI image upscaler that I published as open-source and is free to use.
I also included a link in the repo to my paid app, , which uses the same model but is more user-friendly.
The repo is highly used and drives some traffic to the
1
0
23
Dopamine Sources:
👎Bad:
• Junk food
• Endlessly scrolling feeds
• Netflix
• Video games
• Drugs
👍Good:
• Ice baths
• Workouts
• Meditation
• Cold showers
• Walks
4
0
17
💸 $530 per conversion on
@GoogleAds
– is it just me, or is something off? 🤔
I'm offering a $29 subscription or a $7 one-time payment for my AI Upscaler App. Normally, I have a 1% conversion rate, but with ads, it dropped to 0.007%! 😬
I’m targeting English-speaking countries
9
0
19
Sharing a free version of my app works great to get new users
People are happy to engage with the post bc it's free, and so it reaches new potential users who want something more convenient
✨ Free ComfyUI Workflow for Clarity Upscaler!
Finally, we have a fully ComfyUI workflow that runs locally on your own GPU and uses the same technology as Clarity
To use it, just download the free-workflow.json file from the ComfyUI-ClarityAI repo
😀 Ad: If you're looking for
22
100
684
3
0
18
@atouchofart21
No, it is fully open-source and free to use.
You can run it locally with a1111, which is free:
Or deploy a cog model (more advanced):
0
0
18
Finally, a simple analytics tool that shows where the revenue is coming from.
I'M OVERTHINKING
I'M OVERTHINKING
I'M OV...
I'M LAUNCHING AN EARLY VERSION OF MY NEW SAAS:
📈 — Grow your startup with data (not guesses).
It's an analytics tool for founders to:
• Analyze your traffic
• Find your best marketing channels
• Decide
162
15
745
3
0
18