fal

If you’re at #CVPR in Denver this week, come hang Friday night. 🤙 Andreessen Horowitz is co-hosting a happy hour with a group of companies we’re lucky to work with that are building across multimodal AI: image/video generation, vision systems, spatial intelligence, audio, vector graphics, inference, document agents, and a few things that don’t fit neatly into one bucket: Black Forest Labs - Frontier AI lab for visual intelligence. fal - Generative media platform for developers. Ideogram - Generates stunning images from creative ideas. krea.ai - Creative AI suite for images, video, and 3D. Luma - Creative agents that make you prolific. Mirelo - AI audio models for videos. Overshoot (YC W26) - Vision applications that see and act in real time. Phota Labs - Personal model for visual AI. QuiverAI - Frontier AI for vector graphics. Reducto - Agentic platform for production document processing. Shizuku AI - AI companion from morning to night. World Labs - Spatial intelligence for the 3D world. Should be a fun room: engineers, researchers, and founders. RSVP here: https://lnkd.in/g8JqjPRd

Some jaw-dropping demos last night at our 3D GenMedia Meetup. We brought together 3D frontier builders exploring where AI-generated materials, virtual worlds, animation, and 3D tooling are headed. A few highlights: - Matt Workman demoed his project built with fal Patina and World Labs - Benjamin Paine walked through using Patina inside Blender for rapid rendering - Blendi Bylygbashi showed how he combines a voxelizer with fal APIs to design 3D assets for video games - Ian Curtis demoned World Labs' Marble, which allows you to generate fully navigable 3D worlds from a text prompt, image, video, or rough 3D layout. The throughline: 3D generation is moving from novelty to actual production pipeline, and the builders pushing that frontier are doing it on top of fal. Thanks to Kindred Ventures for hosting us and to everyone who shared their work!

The AI x Hollywood conversation has finally moved past "will it replace artists?" Our CTO, Gorkem Yurtseven was on the AI on the Lot panel last week, and a few things stood out: → Same tools, different artist, different output. Creativity isn't something AI flattens, but enables and amplifies. → "Cinematic storytelling at scale" is an infrastructure problem as much as a creative one. Getting capacity to the right people requires the whole stack: inference, compute, partnerships. That's what we're continuing to build. → Compute is finite. This part of the conversation gets skipped over too often. As AI amplifies creativity, access becomes the real question and the community needs to keep investing together. → Our partnerships with Prime Video & Amazon MGM Studios, Project Nara, and Amazon Web Services (AWS) are about networked enablement, not exclusivity, and enabling creatives to create at scale. There's a real shift happening in the dialogue on what AI even means right now and we're excited to have been part of the conversation. Thank you to Samira Panah Bakhtiar, Albert Cheng from AWS, Caroline Ingeborn from Luma Video AI .

Ideogram 4.0 is now live on fal! SOTA open weights model for realism, text rendering, and artistic generation Typography that reads clean on posters, packaging, and brand work Lightning fast with multiple quality options for every workflow

fal is an official launch partner for OpenAI's new role-specific plugins in Codex, going live today. The bigger shift worth flagging: agents are becoming the surface where creative work actually happens, not just where it gets discussed. With fal in Codex, teams can bring models, assets, and workflows from fal directly into Codex sessions, so analysis, prototyping, creative review, and reporting stay in one place instead of fragmenting across tabs and tools. For creative production teams, that closes the gap between a brief and a useful output. For organizations standardizing on Codex for role-specific workflows, it makes fal's generative infrastructure available exactly where the work is being done.

🚨 LTX 2.3 Quality is now live on fal! 🎬 New Higher Quality LTX 2.3 video generation ⚡ Text, image, audio & reference-video workflows 🎛️ LoRA variants for every endpoint 🌈 HDR with EXR frames + MP4 preview Try It here https://lnkd.in/gXFJSXXf

fal Assets is live. As teams scale AI generation in production, output sprawls. Different modalities land in different drives, generations get re-run because no one can find the original, and lifecycle policies drift across tools and accounts. fal Assets is the consolidation layer. Every image, video, audio file, and 3D model you create on fal lives in one searchable library. Semantic search works across modalities, so a text query can surface a video, or an image can surface a related render. Assets are immediately available as inputs to Sandbox, Playground, or Workflows with no re-uploading.

We're on Redpoint's 2026 InfraRed 100. The pace of AI is accelerating every day and that means infrastructure that can evolved with it. Thank you to Redpoint for the recognition and to all the companies, developers, creators, and builders we get to power every day.

KREA 2 is live on fal. It's KREA's first foundation image model, trained from scratch, and built around a real problem in image generation. Most models force a tradeoff between aesthetic quality and creative control. You can usually get one, rarely both. KREA 2 is designed to hold style, mood, and creative direction consistently across a workflow rather than a single generation. For teams building creative tools (moodboard-driven generation, brand-consistent output, art-directed pipelines), that makes it a stronger primitive than a prompt-tuned model fighting its own aesthetic defaults.

Resonating creativity has always driven industries forward and what Creatify AI is building is an example of how advertising keeps evolving and shifting. We're proud to be the engine behind the team.