Rayan
635 posts
Rayan
@AskRayan
Cofounder CEO @PrunaAI
Paris, France Katılım Temmuz 2021
1.3K Takip Edilen334 Takipçiler
Rayan retweetledi
you'll find inference.sh cli in this @NousResearch Hermes Agent update, here's a quick demo of what it can do with @PrunaAI p- models quickly. ask your Hermes Agent about inference cli 🪽
Nous Research@NousResearch
Hermes Agent v0.3.0 ☤ 248 PRs. 15 contributors. 5 days. • Real-time streaming across CLI and all platforms • First-class plugin architecture, package and share tools+commands+skills • /browser connect to live Chrome via CDP • @vercel AI Gateway model provider • @browser_use browser tool provider • VS Code, Zed, and JetBrains integration • Voice mode with local Whisper • PII redaction everywhere 9 new skills. 50+ bug fixes. Much more in the full changelog.
English
Rayan retweetledi
Rayan retweetledi
all @PrunaAI models are now available: app.inference.sh/apps/pruna
Pruna optimizes AI models for speed without quality loss. Now available:
• P-Image & P-Video - their flagship fast generation
• Optimized FLUX, WAN, Qwen, Z-Image
English
Rayan retweetledi
P-video just got an update! Same price, same speed, better ... 👇
- Better character consistency
- Better sound + audio in draft mode
- Better quality overall
🎠 Try our free demos: dub.sh/sEN7i8c
💜 Sign up for our events: dub.sh/BDmCNBx
🎉 Find all of our models: dub.sh/OP90LS6
⭐️ Star our OSS GitHub: dub.sh/ZA6N6YS
English
Rayan retweetledi
🚀 Release 0.3.2: We’ve added some new algorithms to Pruna OSS!
- Recoverers
- Compilers
- Decoders
- And many, many more…
📚 Check out the docs: dub.sh/Ht2ryVg
⭐️ Star our OSS GitHub: dub.sh/wqToxlX
💜 Sign up for our events: dub.sh/nRlyHuU
English
Rayan retweetledi
P-Video is the fastest and cheapest video model covered by Artificial Analysis, capable of generating a 720p 5-second video in ~10 seconds!
P-Video is the latest video model from @prunaAI, ranking #41 in Text to Video and #42 in Image to Video in the Artificial Analysis Video Arena, comparable to Seedance 1.0 Mini and Wan 2.2 A14B respectively.
At $1.20/min for 720p and $2.40/min for 1080p, P-Video is significantly cheaper than LTX-2 Fast ($2.40/min) and MiniMax's Hailuo 2.3 ($2.80/min) while maintaining faster generation speeds.
P-Video is accessible via @prunaAI's API, as well as @Replicate, @Runware, @Cloudflare, @runpod, @WiroAI, and @Qubrid_AI.
English
Rayan retweetledi
Create an avatar talking, a social ad video, or an animated product image in seconds.
Use simple Workflows with P-Video and P-Image.
- Step 1: Enter a prompt
- Step 2: Get/edit an image
- Step 3: Generate a video in seconds with dialogue, motion, or lookbook shots.
🔗 Try the demo: dub.sh/IHLnuD0
💜 Sign up for our events: dub.sh/Ut89wKZ
🎉Find all of our models: dub.sh/Yn5VWwg
⭐️ Star our OSS GitHub: dub.sh/quN1ETR
English
Seedance 2.0 is honestly the best video AI out there right now. The quality is insane.
But the pricing shows what it probably took to get there. Massive compute, huge datasets, and a lot of brute force scaling.
The real question is whether OpenAI or Google can deliver similar or better video models at a much lower cost. Their scale, especially OpenAI’s, simply doesn’t allow video generation to stay this expensive.
English
cafe compute is scaling up and we need your help deciding where to go next
Cafe Compute is the first, regular late-night coffeeshop for engineers, founders, writers...(anyone) to co-work or hangout late at night.
if you want to cohost or bring cafe compute to your city, comment below 👇
English
Rayan retweetledi
As the cheapest and fastest video model, P-Video strongly sits on the efficiency-quality Pareto frontier
The best speed-cost-to-quality trade-off among video models with public APIs. $1.20/min (Artificial Analysis) vs $4–30 for competitors. ELO 1171 (text-to-video) and 1197 (image-to-video) on arena.ai — ahead of LTX-2, Hunyuan, and in the mix with Hailuo and Seedance.
Sources: @arena · @ArtificialAnlys · @replicate
🔗 Try it: dub.sh/IovIOf9
💜 Sign up for our events: dub.sh/Rh5lVZu
🎉Find all of our models: dub.sh/3YTzAMg
⭐️ Star our OSS GitHub: dub.sh/tLJc2yS
English
@JustinLin610 You did amazing work with the qwen team! A lot of fans back here in France & Germany, come say hi to the community if you've got time for a break :)
We'd be happy to host you with @PrunaAI
English
Rayan retweetledi
P-video is on Prodia! 🚀
* Use draft mode for ultra-fast video generation (5s to get 10s 720p video)
* only $0.025 per 5s of video
* Integrated audio
* Built for real-time creative workflows
* Perfect for ideation: draft → refine → publish
Free gens until Monday 9am CET 🥳
English
🔓 When we say open and 𝐜𝐨𝐧𝐭𝐫𝐨𝐥𝐥𝐞𝐝, we really mean it!
Today we’re releasing a research preview of our next‑gen models - with direct, disentangled control of visuals via both textual and 𝒏𝒖𝒎𝒆𝒓𝒊𝒄 𝒗𝒂𝒍𝒖𝒆𝒔 𝒊𝒏 𝒕𝒉𝒆 𝑽𝑮𝑳.
📄 Paper: arxiv.org/abs/2602.20672
🧪 Demo in Bria AI lab: console.bria.ai/labs/fibo-bbq
Ron Mokady@MokadyRon
Research Preview: Fibo BBQ: Bounding Box & Qolor Control in Large-Scale Text-to-Image Models Text prompts are a terrible UI for precision It’s much more intuitive to drag objects into place or use a color picker than to write “put it 20% left, make it teal, slightly bigger…” Fibo BBQ demonstrates that we can train large-scale T2I models with numeric parameters (e.g., positions / boxes / colors) as part of a structured caption, at scale. This is a step toward richer controllability: more “knobs” beyond text, and new UI paradigms that feel like design tools, not prompt engineering. [Demo and Technical Paper below]
English
This guy grew a SaaS to $750K MRR.
Then he built a new one using the exact same playbook.
This time he hit $1M ARR in 7 months.
Meet @LoicBerthelot:
English
Another weekend, another SaaS built around Replicate's amazing models.
This time the amazing P-Video by Pruna AI.
It's incredible at making photos come to live. Relive your memories.
Video is not sped up or edited in any way.
English
@aifilmmaker Try it with P-Video from @PrunaAI, not as high quality gens but runs in less than 10s so you can iterate very fast, especially early in your creative process, which drives up quality again ! :)
English
I created this screen test to prove to the investors that AI is good enough to make an entire feature film! Unfortunately, Seedance 2.0 got nuked half way through making this, so I had to finish it with Veo and Kling which feels like going backwards 😅 still came out okay, definitely proves the point!
English
@deepwhitman @PrunaAI this is awesome, we might build more stuff for music videos then
English
