The Pruna Build Program is open for teams shipping real AI products. Start with $300 in free credits after signup, then unlock up to $3900 total as your usage grows. No heavy application: add billing info, build, and unlock the next stage when you hit usage milestones. Built for SaaS, API, and infra teams who care about cost, speed, and efficiency. Applications are reviewed monthly (third week of each month). Read how it works: https://lnkd.in/euTDhDFD Apply in 2 minutes: https://lnkd.in/exzAa8yN 🌐 Website: https://www.pruna.ai/ 📚 Documentation: https://docs.pruna.ai/ 🧩 API: https://lnkd.in/e2cmPdzg
Pruna AI
Technologie, Information und Internet
München, Bavaria 6.429 Follower:innen
Making AI accessible & sustainable
Info
At Pruna AI, we build high-performance AI models that combine speed, quality, and specialization for key use cases. We work hand-in-hand with the open-source community through our open packages, models, courses, and educational materials — empowering developers to make their models faster, smaller, cheaper, and greener. 🌱 Our vision is to make AI accessible and sustainable for everyone!
- Website
-
https://www.pruna.ai/
Externer Link zu Pruna AI
- Branche
- Technologie, Information und Internet
- Größe
- 11–50 Beschäftigte
- Hauptsitz
- München, Bavaria
- Art
- Privatunternehmen
- Gegründet
- 2023
Orte
-
Primär
Wegbeschreibung
Freddie-Mercury-Str. 5
München, Bavaria 80797, DE
-
Wegbeschreibung
Paris, FR
Beschäftigte von Pruna AI
Updates
-
With Pruna Skills you can create a full product launch reel from a single agent chat. Install the skills once, start a new chat, and describe the visual flow you want. The agent picks image, video, and audio tools, drafts prompts, calls the Pruna API, and can hand off motion assembly to HyperFrames when you need HTML-to-MP4 polish. Example prompt: > Create a 60s launch reel for Pruna Skills. Hero stills with P-Image-Ideogram, a talking-head intro with P-Video-Avatar, kinetic type for install steps, and a CTA frame. 16:9, brand orange and charcoal. Install: export PRUNA_API_KEY="your_key" npx skills add PrunaAI/pruna-skills@pruna -y 👉 Free Playground: https://buff.ly/qmK81B2 📚 Skills library: https://buff.ly/GySQF1u 🧩 API: https://buff.ly/oo3Lf33
-
Season launches and event drops often need dozens of key art variants, and manual production becomes the bottleneck. With P-Image-Ideogram on Pruna, you can generate cinematic game key art from prompts in four thinking modes, from $0.003 per 1K image at about 0.4s on the fastest tier. What's the setup? 👀 - Model: P-Image-Ideogram - Input prompt: Cinematic key art for your game world, hero, or event drop (style, lighting, composition) - thinking: very low for volume drafts and A/B tests, high for brand-critical heroes More heroes, skins, and campaigns from a brief, not a full production cycle. 👉 Free Playground: https://buff.ly/Y8AIFgm 📚 Documentation: https://buff.ly/oo3Lf33 🧩 API: https://buff.ly/0zfhDKl
-
If every location or franchise has different photos of their dishes, your menu looks inconsistent on delivery apps and socials. With P-Image-Edit on Pruna, you can standardize background, plating, and lighting for one menu look the same across stores, apps, and websites. With P-Image-Edit, expect about 0.9s per edit at $0.010 per run. What's the setup? 👀 - Model: P-Image-Edit - Input image: The photo whose background you want to change - Input prompt: REPLACE the background and plate by a neutral studio set. KEEP the exact same dish from image 1. Ship cleaner listings and campaigns without reshooting. 👉 Free Playground: https://buff.ly/OYJ4Hb3 📚 Documentation: https://buff.ly/bjJJfeh 🧩 API: https://buff.ly/sl6SDnO
-
You already have strong characters, but turning them into presenters for trailers, updates, or community content usually takes animation work. With P-Video-Avatar on Pruna, you give those characters a voice from art plus a script or audio clip, without building a full rig first. About 1.83s per 1s of video at $0.025 per second (720p). What's the setup? 👀 - Model: P-Video-Avatar - Input image: Character art or a portrait with a clear face - Input prompt: Your script or audio, plus framing (e.g. medium close-up, subtle head motion, match the art style) Ship presenter clips for trailers and updates without a new animation pass. 👉 Free Playground: https://buff.ly/PkxtLMr 📚 Documentation: https://buff.ly/bjJJfeh 🧩 API: https://buff.ly/sl6SDnO
-
Introducing the 𝗣𝗿𝘂𝗻𝗮 𝗖𝗿𝗲𝗮𝘁𝗶𝘃𝗲 𝗣𝗿𝗼𝗴𝗿𝗮𝗺 ✨ A new program designed to help creators bring ambitious ideas to life without being limited by time or cost. - 𝗪𝗵𝗮𝘁’𝘀 𝗶𝘁 𝗮𝗯𝗼𝘂𝘁? A space for AI artists and creators to experiment with faster, more affordable, and controllable models. - 𝗪𝗵𝗼’𝘀 𝗶𝘁 𝗳𝗼𝗿? The program is open to artists, filmmakers, animators, designers, developers, creative technologists, and AI-native creators. - 𝗪𝗵𝗮𝘁 𝗱𝗼 𝘆𝗼𝘂 𝗴𝗲𝘁? Become an 𝗘𝘅𝗽𝗹𝗼𝗿𝗲𝗿 or a 𝗙𝗲𝗹𝗹𝗼𝘄 and get supported with monthly credits, early model access, rewarded campaigns, and opportunities to be featured. No need for thousands of followers. Just original ideas. Your work. Your style. Your creative freedom. Read more information and apply here: https://buff.ly/9xNoyxr
-
P-Image-Ideogram now works with Pruna Skills 🚀 - Install: npx skills add PrunaAI/pruna-skills@p-image-ideogram -y - Then ask your agent for image workflows Example prompt: > Need a wide 16:9 hero for a sneaker drop. Guy walking through a neon warehouse, hype streetwear ad vibe. Saying something motivational. 👉 Free Playground: https://buff.ly/QmnA5P5 📚 Skills library: https://buff.ly/C09F3qA 🧩 API: https://buff.ly/bjJJfeh
-
Pruna AI hat dies direkt geteilt
Today we are releasing something great! We partnered up with Caveman and Ponytail (Dietrich Gebert) So.....Less talk. Less code. Same answer. We're shipping GreenPT compression models, and we're the first AI provider in the world to supply this natively on the API endpoint level. Compression happens before the response ever reaches you. Same model. Same price per token. Same CO2 per token. Up to 70% fewer output tokens, because the answer comes back shorter. How it works Choose a family by model id, no extra parameters, no SDK gymnastics: 🌱 glm-5.2-caveman, terser prose. Strips filler, keeps every technical token intact. (-22% to -87%) 🌱 glm-5.2-ponytail, less generated code. Your agent thinks like the laziest senior dev in the room. (-22% to -94%) Each comes in -lite, default and -ultra intensities. Swap the model id in your existing OpenAI-compatible request and you're done. ```json { "model": "glm-5.2-ponytail", "messages": [{ "role": "user", "content": "Add retry with backoff to this fetch call." }] } ``` Works in OpenCode, and every coding app that speaks the OpenAI API. The compression is baked into the endpoint itself. The story behind the names We didn't invent these rulesets. Two brilliant open-source builders did: 🪨 Caveman, by Julius Brussee, makes your agent talk like a caveman: same answers, roughly 65% fewer output tokens, code and errors byte-for-byte exact. A March 2026 paper found brevity can actually improve accuracy by around 26 points. 94.5k stars on GitHub. 🎀 Ponytail, by Dietrich Gebert, makes your agent code like the senior dev who's been there longer than the version control. YAGNI first, stdlib before custom code, never compromise on validation or security. Around 54% less code, up to 94% on over-builds, 100% safe. 92k stars. Why it matters Same answer, delivered with less energy. Fewer output tokens means less compute, less CO2 and a faster response, at no extra cost to you. The saving comes entirely from generating fewer tokens, and we pass all of it through. About the video Yes, we made a video about efficient AI using AI. We get that that's funny. We couldn't resist: the meeting between a caveman and a senior dev practically begs for a video, and once we saw the cinematic possibilities we caved (sorry). We did try to do it responsibly, with Pruna AI AI's P-Video, three to six times more energy-efficient than comparable models. Less talk. Less code. Same answer. Same planet. 💚 Docs and links in the comments.
-
All benchmarks agree: P-Image-Ideogram break the efficiency-quality Pareto frontiers. - Human & AI judges agree on P-Image-Ideogram dominating quality-speed & quality-price Pareto frontiers. - On key categories like safety, fairness, quality, realism, or details, P-image-Ideogram reaches top #1-#3 rankings against 35 models. - P-Image-Ideogram strictly beats 25 models on both price & quality or speed & quality. This is validated across sources by Design Arena, Datapoint AI, Rapidata, and internal benchmarks! 👉 Playground: https://buff.ly/QmnA5P5 📚 Model page: https://buff.ly/XRaKSb0 🧩 API: https://buff.ly/bjJJfeh
-
-
Pruna AI hat dies direkt geteilt
BREAKING: p-image-ideogram by Pruna AI and Ideogram has been released and establishes new Pareto frontiers for both Preference vs. Price and Preference vs. Speed. At 1K resolution, the very_low reasoning variant costs just 0.3¢ per image, while low costs 0.75¢ per image. Both generate images in around 3 seconds - roughly 8.6× faster than the Design Arena average. This is a major step forward in making high-quality image generation faster and more affordable. Huge congratulations to both teams!
-