llms.txt
Compare with other sites' llms.txt files →# Pruna AI - Performance Models
> Pruna AI provides performance AI models and optimization tools that help teams run image and video generation faster, cheaper, more sustainable, and more efficiently.
## Performance Models
- [P-Image](https://www.pruna.ai/p-image): Fast text-to-image model designed for high-quality generation, strong prompt adherence, and reliable text rendering.
- [P-Image-Edit](https://www.pruna.ai/p-image-edit): Image editing model built for fast, high-quality edits with strong instruction following, multi-image editing capabilities, and precise text rendering.
- [P-Video](https://www.pruna.ai/p-video): Video generation model focused on fast, high-quality output, multimodal support (text, image, audio), and draft mode for rapid iteration.
- [All Models](https://www.pruna.ai/all-models): Complete catalog of Pruna and partner models, including formats, inference times, and pricing details.
## Endpoints
- [Pruna API Portal](https://dashboard.pruna.ai/login): Get started with Pruna API, create an account, and start using the API.
- [Quickstart](https://docs.api.pruna.ai/guides/quickstart): Quickstart guide for integrating Pruna APIs into applications and workflows.
- [Pruna API Reference](https://docs.api.pruna.ai/): API reference documentation for the Pruna API.
- [Prompting documentation](https://docs.pruna.ai/en/stable/docs_pruna_endpoints/index.html): Prompting guides for image generation, image editing, and video generation.
### Demos
- [Performance Model Demos](https://demo.pruna.ai/): Main landing page for Pruna’s public demos to get started with Pruna API. It includes llms, image generation, image editing, video generation and LoRAs for different use cases.
- [Virtual Try-On Haul](https://demo.pruna.ai/try-on-haul): Generate people and clothing items, then preview how they look together.
- [Print Studio](https://demo.pruna.ai/print-studio): AI image creation and transformation demo studio.
- [Pitch Mood Board Generator](https://demo.pruna.ai/pitch-mood-board): Generate campaign and mood board concepts in real time.
- [Lightning Documents](https://demo.pruna.ai/lightning-documents): Demo page for Pruna’s document-oriented generation workflow.
- [Ecommerce Ad Studio](https://demo.pruna.ai/ecommerce-ad-studio): Generate and edit ecommerce marketing creatives.
- [Chat Image Editor](https://demo.pruna.ai/chat-image-edit): Edit images through natural-language instructions.
- [Image Merger](https://demo.pruna.ai/image-merger): Upload and combine multiple images into one AI-generated composition.
- [Virtual Staging](https://demo.pruna.ai/virtual-staging): Stage interior spaces with AI-generated furniture and decor.
- [Stock Photos Generator](https://demo.pruna.ai/stock-photos-generator): Generate stock-style marketing and commercial images from prompts.
- [AI Photoshop Brush](https://demo.pruna.ai/photoshop): Brush-based AI image editing demo.
### Pricing
- [Pricing](https://www.pruna.ai/pricing): Commercial overview for public API access and enterprise options.
- [API Pricing for all available models](https://docs.api.pruna.ai/guides/models): Public pricing for image, video, editing, and related model usage.
## FAQ
- [FAQ](https://docs.api.pruna.ai/faq): Frequently asked questions covering latency, sync vs async behavior, polling, uploads, public image URLs, rate limits, resolution handling, seeds, LoRA support, and safety settings.
- [Performance Issues](https://docs.api.pruna.ai/faq#q-my-generations-feel-slow-whats-going-on): Explains common causes of slow generations, including server distance and unnecessary extra API calls.
- [Sync vs Async & Timeouts](https://docs.api.pruna.ai/faq#sync-vs-async-timeouts): Describes how `try-sync: true` works, when to prefer async jobs, and how to avoid timeouts on long-running requests.
- [Input Images – Base64](https://docs.api.pruna.ai/faq#input-images-base64): Clarifies that Base64 uploads are not supported and recommends either Pruna’s upload endpoint or a public HTTPS image URL.
- [Rate Limits](https://docs.api.pruna.ai/faq#rate-limits): Explains per-account request limits and notes that higher throughput can be enabled on request.
- [Seed](https://docs.api.pruna.ai/faq#seed): Notes that the API supports seeds for reproducibility or variation.
- [Safety / NSFW](https://docs.api.pruna.ai/faq#safety-nsfw): Documents the NSFW filter as a configurable input parameter rather than a single enforced global policy.
## Open Source
- [Pruna GitHub Repository](https://github.com/PrunaAI/pruna): Source code, examples, and project documentation for the open-source optimization framework built for developers, enabling developers to deliver faster, more efficient models with minimal overhead. It provides a comprehensive suite of compression algorithms, including caching, quantization, pruning, distillation, and compilation techniques.
- [Install Pruna](https://docs.pruna.ai/en/stable/setup/install.html): Installation requirements and environment setup for the `pruna` package.
- [Algorithms Overview](https://docs.pruna.ai/en/stable/compression.html): Overview of supported compression and optimization methods, including quantization, pruning, distillation, caching, and compilation workflows.
- [Evaluation](https://docs.pruna.ai/en/stable/docs_pruna/evaluate.html): Evaluation metrics for image, video, and LLM models.
- [Tutorials](https://docs.pruna.ai/en/stable/docs_pruna/tutorials/index.html): Hands-on tutorials for compressing, evaluating, and deploying optimized models.
- [Pruna OSS documentation](https://docs.pruna.ai/en/stable/): Guides for installing, configuring, and using the open-source `pruna` optimization framework.
## AdditionalResources
- [Blog](https://www.pruna.ai/blog): Articles, guides, and case studies on model optimization, deployment, and generative AI performance.
- [Events](https://luma.com/pruna-ai?k=c): Information on upcoming online or in-person events with Pruna AI and other industry experts.
- [Infer Bench](https://huggingface.co/spaces/PrunaAI/InferBench): Benchmarking tool for image and video generation models, allowing you to compare the performance of different models and compression techniques.
- [Sustainability](https://www.pruna.ai/solution-pages/user-sustainability-officers): Overview of the sustainability challenges, efforts and commitment to the environment.
### Case Studies
- [Accelerating FLUX.2 [flex]: Now Design x3 Faster](https://www.pruna.ai/blog/flux2flex-3-faster): Case study on Pruna’s collaboration with Black Forest Labs to accelerate FLUX.2 [flex], reducing text-to-image latency from about 48.5s to 13.1s per image and image-edit latency from about 104.8s to 27.6s while preserving quality for typography-heavy workflows.
- [How Fimo.ai Built an Image Gen Feature with Pruna.ai](https://www.pruna.ai/blog/p-image-fimo): Case study on how Fimo.ai evaluated production image generation requirements such as latency, consistency, and scalability while building an AI-native CMS experience with Pruna. [oai_citation:1‡pruna.ai](https://www.pruna.ai/blog/p-image-fimo)
- [P-Image and P-Image Edit are live on Scenario](https://www.pruna.ai/blog/p-image-edit-scenario): Partnership case study on Scenario adopting Pruna’s fast, low-cost image generation and editing models, with the CEO describing the result as one of Scenario’s top three favorite models and recommending it to 30+ customers. [oai_citation:2‡pruna.ai](https://www.pruna.ai/blog/p-image-edit-scenario)
- [Flux 2: A 110GB Model Now Running on One H100 and 4.2× Faster!](https://www.pruna.ai/blog/flux2-100-faster): Technical case study on optimizing Flux 2 Dev so it can run on a single H100 with significantly faster inference, while highlighting improvements for photorealism, control, and multi-reference creative workflows. [oai_citation:3‡pruna.ai](https://www.pruna.ai/blog/flux2-100-faster)
- [Inside Story: The Pruna × Replicate Recipe for the Fastest Inference](https://www.pruna.ai/blog/pruna-replicate-recipe-fastest-inference): Partnership case study on how Pruna and Replicate worked together during 2025 to optimize hosted model endpoints, including early Flux and HiDream deployments with measurable speedups and large-scale usage. [oai_citation:4‡pruna.ai](https://www.pruna.ai/blog/pruna-replicate-recipe-fastest-inference)
- [How Pruna Delivered Up to 3.6x Faster Inference for BRIA 3.2](https://www.pruna.ai/blog/4x-faster-bria32): Case study on Pruna’s work with BRIA to optimize BRIA 3.2, delivering runtime configurations with roughly 2× to 3.6× faster inference while preserving image quality validated by BRIA’s visual team. [oai_citation:5‡pruna.ai](https://www.pruna.ai/blog/4x-faster-bria32)
## Company
- [About](https://www.pruna.ai/about): Company background, mission, team, and approach to AI efficiency.
- [Contact Us](https://www.pruna.ai/contact): Contact page for product questions, partnerships, trials, and enterprise discussions.
- [Discord](https://discord.gg/JFQmtFKCjd): Connect with the Pruna community for discussions and support.
### Legal Information
- [Terms](https://docs.api.pruna.ai/terms): Terms of service for the Pruna API.
- [Privacy Notice](https://www.pruna.ai/privacy-notice): Privacy policy for the Pruna AI.
- [Legal Notice](https://www.pruna.ai/legal-notice): Legal notice for the Pruna AI.