# Pixparkle > Pixparkle is a free AI image and video generator with a chat-based interface. Built for content creators, e-commerce sellers, and social media creators who need high-quality visuals without complex tools or form-based workflows. ## Core Capabilities - AI image generation from text descriptions in plain language - AI image generation from an existing image (image-to-image, up to 14 reference images) - Multi-turn editing: refine results through natural conversation without quality loss - High-resolution output up to 4K (3840x2160) for print, packaging, and large-format use - Accurate multilingual text rendering: Chinese, Japanese, Korean (CJK) characters render cleanly at full resolution - Image-to-video generation in the same chat session - Browser-based: no installation required, works on desktop and mobile - Commercial use allowed on all plans ## AI Models - GPT Image 2: precise text rendering and photorealistic output - Nano Banana: fast generation with strong visual quality - Nano Banana 2: fast generation with strong visual quality - Flux: rapid low-cost drafts - Seedream: video generation - Veo: video generation ## Pricing - Free: $0, 10 credits on sign-up, no credit card required, access to all AI models - Starter: $9.9/month, 500 credits, all models, 4K resolution, priority queue - Pro: $29.9/month, 1,600 credits, all models, best credit rate - Premium: $99.9/month, 6,500 credits, all models, lowest cost per image - Annual subscription: saves 17% vs monthly billing ## Target Users - Content creators: thumbnail generation, cover design, iterative visual drafts - E-commerce sellers: Amazon-compliant white-background product shots, lifestyle images, 9-image listing sets - Social media creators: multilingual content, brand visuals, fast iteration ## Differentiation - Chat-based interface: describe, see result, reply to refine. No form fields, no restarting. - Multi-turn editing without quality loss: each message refines the existing image in context - Multi-model access in a single session: switch models without losing conversation history - Image-to-video in the same session: no separate tool required - Accurate CJK text rendering: handles Chinese, Japanese, and Korean where most AI image generators fail - Curated AI image prompt library plus an AI image prompt generator that outputs both natural-language prompts and reusable JSON style templates, cross-model compatible (Nano Banana, GPT Image, Flux, Gemini Image) - Transparent pricing: all four plan tiers listed publicly with credit counts and per-image cost ## Resources - Homepage: https://pixparkle.com - AI Image Prompts (library, generator, JSON style templates): https://pixparkle.com/ai-image-prompts - Blog (long-form guides on JSON prompts, character consistency, model workflows): https://pixparkle.com/blog - What Is a JSON Prompt? (flagship guide on reusable AI character and style systems): https://pixparkle.com/blog/guides/what-is-a-json-prompt - Pricing: https://pixparkle.com/pricing - Changelog: https://pixparkle.com/changelog - Support: https://pixparkle.com/support