Mark Ledford

255 posts

Mark Ledford banner
Mark Ledford

Mark Ledford

@fishmongr

Professional engineer, creative technologist, and sometimes Twitter lurker. CTO, Team Sogni AI ✨

Seattle Katılım Temmuz 2008
290 Takip Edilen475 Takipçiler
Sabitlenmiş Tweet
Mark Ledford
Mark Ledford@fishmongr·
Absolutely floored by the image-to-video results we're getting from the new LTX-2.3 video model that just went live on app.sogni.ai and the Sogni Client SDK over the weekend. This latest Open Source model is so incredible at everything we throw at it, particularly when using our new Qwen3.5 35B A3B powered prompt enhancer. Here's a 🧵 of some of my personal favorite examples with prompts included:
English
7
5
12
548
Ethan Kiddo
Ethan Kiddo@TshirtCamo·
I love the direction @Sogni_Protocol has been taking lately. been waiting for this since long ago. the team is shifting focus toward everyday users, rather than just building a high-level tool strictly for artists and professionals. have been playing with the first version of their chatbot - you can create images, videos, and audio using simple prompts, just like other mainstream bots. still seeing a few bugs here and there, but it’s defo a move in the right direction!
Ethan Kiddo tweet media
English
2
1
4
108
Mark Ledford retweetledi
Sogni.ai
Sogni.ai@Sogni_Protocol·
Say OpenClaw one more time. I dare you. 🤨 Recreating classic movie scenes with LTX-2.3. One scene turns into ten real fast!
English
3
7
29
832
Basit
Basit@soft4211·
@fishmongr Thanks Mark for giving your precious time reading this thread, I am going to check this feature ASAP, Sogni is moving at lighting speed, don't feel bad if someone can't catch up haha You might also like this thread I wrote on the Sogni infra & SDK x.com/i/status/20320…
Basit@soft4211

This is how to use Sogni Protocol Productively (Including SDK): 🧵 I’ve been experimenting with the Sogni ecosystem recently, and the SDK is honestly one of the more interesting tools if you're building around creative AI. Here’s how I use it and what makes it useful 👇

English
1
0
1
28
Mark Ledford
Mark Ledford@fishmongr·
@soft4211 Awesome recap @soft4211 ! We’re constantly adding new features. Try uploading 2 images with the “attach media” + button, then asking it to restyle the first image into the style of the second!
English
1
0
1
122
Basit
Basit@soft4211·
The best thing I’ve found is that I can create anything, including images, videos, and even songs. How? Just chat with the model—it can generate images (text-to-image), turn images into videos (image-to-video), or create videos from text (text-to-video). All you need is an account at app.sogni.ai. Sign up, claim your free 50 SPARKS (daily) signup with my referal and you get a bonus (soft4211), and start generating!
Basit tweet media
English
2
0
1
36
Mark Ledford retweetledi
Sogni.ai
Sogni.ai@Sogni_Protocol·
Invite creators. Earn Spark. Climb the leaderboard. Sogni Ambassador is a referral program that rewards creators for helping grow the Sogni community. Your username becomes your referral code. When someone joins through you and makes their first Spark purchase, you both earn Sparks. You also collect Ambassador Points through your own activity and the activity of people you invite. Those points help determine your place on the leaderboard and your share of the $SOGNI token airdrop. ...and don’t forget, you can also create wild videos with the LTX-2.3 video model using just a single prompt! 🎬 Learn more below 👇
English
9
9
35
845
Mark Ledford
Mark Ledford@fishmongr·
@soft4211 Maybe share it on the Discord? I can't really beta test, have a company to run!
English
1
0
0
9
Basit
Basit@soft4211·
@fishmongr Do you mind trying it now? I just setted an EC2 instance in Cloud AWS, and making the version for the frontend only requests now
English
1
0
0
22
Mark Ledford
Mark Ledford@fishmongr·
Absolutely floored by the image-to-video results we're getting from the new LTX-2.3 video model that just went live on app.sogni.ai and the Sogni Client SDK over the weekend. This latest Open Source model is so incredible at everything we throw at it, particularly when using our new Qwen3.5 35B A3B powered prompt enhancer. Here's a 🧵 of some of my personal favorite examples with prompts included:
English
7
5
12
548
Mark Ledford
Mark Ledford@fishmongr·
@labuanjo haha it gets pretty fun, I've been generating them in the AI Chat feature on restore.sogni.ai. We're going to add that to more apps stat!
English
0
0
1
23
Mark Ledford
Mark Ledford@fishmongr·
Peer-to-peer cloud scale without the Content Policy Violations 🤔
English
0
0
0
27
Mark Ledford
Mark Ledford@fishmongr·
Skull crushingly good ❤️
English
0
0
0
24
Mark Ledford
Mark Ledford@fishmongr·
Render what you want
English
0
0
0
23
Mark Ledford
Mark Ledford@fishmongr·
Cheaper than a cup of coffee ☕️
English
0
0
0
29
Mark Ledford
Mark Ledford@fishmongr·
Man I sure do love hearing about Open Claw
English
1
0
0
43
Mark Ledford
Mark Ledford@fishmongr·
Giving you what you want
English
0
0
0
26
Mark Ledford
Mark Ledford@fishmongr·
Two green felt Muppet puppets face each other against a flat, matte gray background under even, diffused studio lighting. The puppet on the left has smooth white plastic eyes and a soft, fibrous green body; the puppet on the right is draped in a black hooded cloak made of matte fabric that obscures its head, leaving only its green felt snout visible beneath the shadow. Both puppets are positioned upright, facing one another with no motion implied — their expressions are neutral, static, and unchanging. White sans-serif text overlays the top and bottom edges of the frame, contrasting sharply against the muted tones of the scene. The dark muppet frog under the cloak stays completely still, only his mouth moves. He speaks in an extremely soft, barely audible whisper, with heavy, uncomfortable breathing throughout the entire scene. He says: "Do it." Long pause with heavy breathing. "Just one more meme." Another long pause with heavy breathing. "Just one more meme." Long pause with heavy breathing. "Do it." His expression is blank and emotionless, his voice barely audible over the breathing. The atmosphere is quiet, intense, and uncomfortable. The camera is static. The footage remains smooth and stabilised throughout.
English
0
0
0
65
Mark Ledford
Mark Ledford@fishmongr·
Hey Basit! Nice, looking cool! I tried to use it with a Sogni API Key but it threw an error. Looks like it was still trying to hit "localhost:3001/api/generate-song-video" You can have your agent reference github.com/Sogni-AI/sogni… on implementation detail on fully switching to Sogni Client SDK via front-end when the user is logged in. It shouldn't need to make a call to your local server if it's making calls directly to Sogni Client SDK on the front-end. Also if it's making localhost calls it means the ENV Vars for prod endpoints must not be set up during deploy.
English
3
0
0
29