Pixlie
AI video studio: text & image to video, with real control
127 followers
AI video studio: text & image to video, with real control
127 followers
Pixlie is an AI video studio for creators who want control—not a black-box generator. Browse your library, queue jobs in the cloud, and track renders in the app. Same account on pixlys.com and iOS.What makes it different: Text2Video, Image2Video, library & cloud queue Granular creative workflow not just “type and pray” Live on iOS App Store; Google play web at pixlys.com -Built solo by a refugee founder in Helsinki shipped fast, iterated in production
Products used by Pixlie
Explore the tech stack and tools that power Pixlie. See what products Pixlie uses for development, design, marketing, analytics, and more.
Platforms 1
Platforms 1
YC Startup ClassA live, one-hour class by Sam Altman on Platzi
5.0 (1 review)
Also considered:
Startup SchoolStartup School taught us the basics; YC pushed us to talk to users weekly and ship something people pay for—not another wrapper. For Pixlie, that meant nailing creator workflow (prompt → queue → library → export), not chasing feature checklists.
Design & Creative 1
Design & Creative 1
LambdaTurn camera shot into artwork with AI
5.0 (3 reviews)
We needed a GH200-class box with fast filesystem attach not a hobby GPU hour. Lambda let us mount weights once, run renders in minutes, and terminate when idle. For a solo founder, that’s the difference between cool prototype and users get their video tonight.
LLMs 2
LLMs 2
NVIDIAThe official handle for NVIDIA.
5.0 (30 reviews)
Pixlie is a GPU-heavy video product—Wan workflows, ComfyUI, long renders. NVIDIA’s ecosystem (Inception, docs, CUDA-optimized stacks) made it realistic to ship director-level AI video, not just pretty demos. Hugging Face is essential for models, but NVIDIA is the layer that turns those weights into production throughput we can actually sell.
OpenAIAPIs and tools for building AI products
5.0 (777 reviews)
OpenAI models power parts of our creative and ops loop—launch copy, prompt iteration, and tooling glue while we keep heavy video gen on ComfyUI/GPU. Reliable APIs and fast iteration let us stay focused on product video quality, not building another LLM stack.
Engineering & Development 2
Engineering & Development 2
Google Cloud PlatformA suite of cloud computing services by Google
5.0 (188 reviews)
Firebase
Herokue run Pixlie’s web on Firebase Hosting and batch-generate launch creatives with Veo on Google’s stack. Firebase gave us fast global deploy for pixlys.com without babysitting CDN config, and GCP’s video/ML APIs let us produce Product Hunt b-roll at studio quality from prompts—not stock footage. Heroku stayed great for our API worker, but for frontend + generative media pipeline GCP was the one place that scaled with us.
General 1
General 1
nebiusAbout Nebius
5.0 (1 review)
Nebius was our bridge from experiment to EU-friendly inference: dedicated endpoints, Token Factory, and serious H200 capacity when we were still iterating workflows. It let the team ship fast without building our own GPU ops and kept model serving predictable while we moved prod render to Lambda