
Turn Any Media IntoProduction AI Prompts
Counts requests from this browser today. ≈ limits are estimates until a key reaches its real limit.
What is ProPeek?
ProPeek is an open, client-side reverse-engineering suite created for digital artists, videographers, and stock contributors. Instead of manually guessing how complex generative videos or photorealistic stills were constructed, ProPeek analyzes any uploaded video clip or image to reconstruct the exact cinematography, lighting setups, color palettes, and motion dynamics into production-grade prompts. Engineered with a strict privacy-first architecture, ProPeek runs entirely within your browser using hardware-accelerated HTML5 decoding and direct multimodal API connections to Google Gemini and Groq. No user media, prompt history, or secret credentials are ever uploaded or retained on an intermediary server, giving creative professionals uncompromising security, speed, and creative autonomy.
How to Use ProPeek
Follow this five-step workflow to reverse-engineer media into actionable prompts and commercial metadata.
Connect Your API Key
Open API Settings and enter your free Google Gemini or Groq API token. Keys are saved securely in your browser's encrypted localStorage and never leave your machine.
Select Your Mode
Choose between Video to Prompt, Image to Prompt, Metadata Only, Video to Script, or Niche Replicator depending on your production goals.
Drop Your Files
Drag and drop single or batch media files (MP4, MOV, WebM, PNG, JPG). ProPeek's HTML5 canvas instantly samples temporal keyframes locally without waiting for slow server uploads.
Tune Style & Intent
Select Recreate to reconstruct the asset with pixel precision, or Inspire to derive an original scene matching its cinematography without copyright conflicts.
Generate & Export
Hit Analyze to trigger multimodal inference. Copy formatted prompts for your favorite generative engine, or export complete Adobe Stock and Shutterstock CSV spreadsheets with one click.
Who is ProPeek Built For?
Microstock Contributors
Stock videographers and photographers submitting to Adobe Stock, Shutterstock, Pond5, and Freepik. Eliminate hours spent manually typing titles, assigning category IDs, and compiling 25 to 49 ranked, non-spam keywords that pass automated review.
AI Video Directors & Animators
Filmmakers producing clips with Google Veo 3, ByteDance Seedance 2.0, Runway Gen-3, Kling AI, and Luma. Reverse-engineer pacing, camera tracking, and lens mechanics from reference footage to generate coherent cinematic scenes.
Prompt Engineers & Content Teams
Digital agencies and content creators managing faceless YouTube channels, viral TikTok accounts, and brand asset pipelines. Deconstruct viral aesthetic trends into reusable template structures for consistent batch content delivery.
Supported Generative AI Models
ProPeek constructs customized prompting syntax tailored specifically for the leading video and image diffusion architectures.
Google Veo 3
Google's state-of-the-art cinematic video foundation model. ProPeek generates Veo 3 prompts with structured temporal momentum, lighting consistency, and camera framing tokens that preserve subject identity across 1080p and 4k scene generations.
ByteDance Seedance 2.0
High-velocity, hyper-realistic motion generation engine engineered for complex human movement and cinematic action. Our prompt synthesizer formats multi-stage acceleration, camera rotation, and physical obstacle interactions tailored for Seedance.
Midjourney v6.1
The benchmark in photorealism and illustrative rendering. ProPeek formulates prompts including aspect ratios, stylize values, photographic aperture, shutter speeds, and film stock aesthetics without confusing negative prompt contamination.
Black Forest Labs Flux.1
Leading open-weights diffusion model renowned for accurate anatomical typography and intricate spatial compositions. ProPeek constructs natural-language descriptive prompts that maximize Flux's guidance adherence and granular texture fidelity.
Reverse-Engineering in Action: Input vs. Output Prompt
See how a simple raw stock video clip is deconstructed into a comprehensive production prompt and metadata record.
Espresso Pouring in Morning Sunlight
A barista pulling an espresso shot into a heavy ceramic cup. Golden morning sunlight beams through a cafe window, illuminating airborne steam swirls and glossy crema micro-bubbles. Slow smooth pan left to right.
Cinematic Generation Directive
Cinematic slow-motion close-up of a double espresso stream cascading from a brass portafilter into a textured ceramic cup. Thick, hazelnut-hued crema forming swirling velvet micro-patterns. Warm golden morning sunlight streaming at a 45-degree angle through rustic cafe window glass, backlighting delicate wisps of rising steam. Smooth horizontal dolly tracking shot from left to right, shallow 50mm f/1.8 depth of field, photorealistic lighting, 8k resolution, 60fps.
Join Our Official Networks & Grow With Us
ProPeek is a free tool by Elite Squad, built by Ali Mr Nobody. Follow the official Elite Squad channel below for every new Elite Squad tool release, or join Ali Mr Nobody's own channel for AI prompting and automation guides.
Elite Squad
AI Creative Suite & Hub- Official Tools Updates
- Daily Workflow Guides
- Exclusive Creator Access
- Latest Tools Access
Ali Mr Nobody
AI & Automation Insights- Advanced AI Prompting
- Tool Demos & Guides
- Direct Community Support
- Automation Strategies
Hamza Creator Academy
Creator Growth & Skills- Masterclasses & Tutorials
- Monetization Strategies
- Step-by-Step Roadmaps
- Creator Growth Mentorship
Elite Tools
Elite Squad AI Tools Hub- Elite Squad Tools Access
- Early Beta & New Tool Access
- Instant Tool Releases
- Continuous Feature Updates
AI Tools & Code Hub
Code & AI Utilities- Curated Free AI Tools
- Code Snippets & Scripts
- Latest Tech Updates
- Daily Pro AI Utilities
Mughal.Dev
Dev Tools & Free Resources- Free Premium Themes & Tools
- Daily Fresh Resource Drops
- Direct Ad-Free Downloads
- GitHub Repos & AI Prompts
Frequently Asked Questions
Reverse-engineering prompts, stock metadata, and prompt variations — how ProPeek's modes actually work.
What does ProPeek actually do?
ProPeek reverse-engineers any stock video or image into a production-ready AI generation prompt for tools like Veo 3, Seedance 2.0, Midjourney and Flux, and can generate Adobe Stock or Shutterstock-ready title, keyword and category metadata from the same file. Everything runs in your browser — nothing is uploaded anywhere except the AI provider you choose.
How does ProPeek work?
Drop a video or image, pick a style — or keep it exact with Recreate — and ProPeek's AI reads it using your own free Gemini or Groq key, entirely in your browser. A few seconds later you get a production-ready prompt for Veo 3, Seedance, Midjourney or Flux, plus Adobe Stock-ready title, keywords and category if you need them. It's a free AI prompt generator: nothing is uploaded to a ProPeek server at any point.
What's the difference between Video to Prompt, Image to Prompt, Metadata Only, and Video to Script?
Video to Prompt turns a clip into a generation prompt (Veo 3, Seedance, Muse) and Image to Prompt turns a still into one (Midjourney, Flux), with Same Image, Thumbnail and Character Sheet presets. Metadata Only skips the creative prompt and returns just the stock upload fields — title, description, keywords and category — ready for the Adobe Stock or Shutterstock CSV, at about half the output of a full analysis. Video to Script transcribes a video into a ready-to-use, timestamped script.
What is Prompt Lab?
Prompt Lab works from typed text instead of a file, with three tools: Enhance turns a rough idea into a full generator prompt, Prompt → Metadata infers stock upload fields from a prompt you already have, and Variations expands one prompt into a matching set — you choose how many variations.
What's the difference between Recreate and Inspire?
Recreate reverse-engineers the exact asset you dropped into a prompt that rebuilds it. Inspire studies only its style — mood, palette, lighting, camera language — and writes a prompt for a brand-new scene in that style, never naming an artist or copying the original. This matters for Adobe's AI-submission guideline, which forbids a prompt "generated with reference to other artist(s)": a Recreate prompt can trip that rule, an Inspire prompt cannot.
Is my video or image uploaded anywhere?
No. Every video and image is processed entirely in your browser. The only network call ProPeek makes with your media is to whichever AI provider you've added a key for — Gemini, Groq, OpenRouter, or a local Ollama install. Nothing is stored on or sent to a ProPeek server, because there isn't one.
Which AI providers does it support, and which one reads real motion?
Google Gemini, Groq, OpenRouter, and a fully offline Ollama setup. Only Gemini reads the actual uploaded video file — every other provider works from sampled keyframes and infers motion from the differences between them. For image prompts, all four are equally capable.
Can I process many files at once?
Yes. Drop or select multiple videos or images and they queue up; ProPeek runs them strictly one at a time to stay under free-tier per-minute rate limits, and every result lands in your library as its own card you can copy, edit or export.
Does it export a file I can upload straight to Adobe Stock or Shutterstock?
Yes. Export CSV writes Adobe Stock's exact Filename/Title/Keywords/Category/Releases format or Shutterstock's own format, with titles sanitised, keywords deduplicated and capped, and categories mapped onto each platform's official list. Anything that would get a row rejected is flagged before you download, not after you upload.
How many variations can I generate from one prompt?
Between 2 and 10 per prompt, adjustable with a slider in Prompt Lab's Variations tool.
Is ProPeek free to use?
Yes. ProPeek itself is free and open — you bring your own API key from a provider with a free tier (Gemini and Groq both offer one), and ProPeek never asks for payment or an account.
How does ProPeek's multi-key rotation work and why is it needed?
Free AI tiers (like Google Gemini's 15 requests-per-minute quota or Groq's token limits) can get throttled during batch processing. ProPeek lets you enter up to 3 Gemini keys and 2 Groq keys simultaneously. When analyzing batches, our client-side request engine automatically rotates between your active keys round-robin, preventing HTTP 429 rate-limit errors and ensuring uninterrupted high-speed batch processing.
Are ProPeek prompts and metadata safe for commercial Adobe Stock AI guidelines?
Yes. ProPeek's prompt sanitization pipeline and Inspire mode strictly follow Adobe Stock's Generative AI Contributor Guidelines. It strips copyrighted artist names, trademarks, and disallowed spam phrases. Furthermore, our metadata engine compiles exactly 25 to 49 unique, deduplicated keywords and maps them to Adobe Stock's official category IDs to ensure fast moderation approval.
Can I run ProPeek on low-spec laptops or mobile devices?
Yes. ProPeek is lightweight and utilizes hardware-accelerated HTML5 video decoding directly in the browser canvas. Instead of loading entire video files into memory, it sequentially samples frame snapshots and immediately discards decoded buffer data, maintaining a tiny memory footprint (<50MB) that runs smoothly on standard laptops, Chromebooks, and smartphones.
How does in-browser keyframe extraction preserve privacy?
Unlike cloud-based conversion tools that require you to upload your raw footage to their remote servers, ProPeek operates entirely client-side. Keyframes are extracted locally inside your browser runtime. The only network call made with those frame snapshots goes directly over secure HTTPS to the official AI provider (Google or Groq) using your own personal API key. No intermediate server or third party ever sees your content.