Vision GPT Extension

Analyze screenshots with GPT‑5.6 and other vision-capable OpenAI models—directly with your own API key. VisionGPT Extension is a privacy-focused, BYOK screenshot analysis tool for developers, designers, product teams, QA engineers, and anyone working with sensitive content. Capture any area of a webpage, choose a prompt, and get an AI-powered analysis in seconds. 🔐 Privacy-first and direct - Bring your own OpenAI API key - No VisionGPT server, proxy, account, or license validation - API requests go directly from the extension to OpenAI - Screenshots are stored locally in your browser - Screenshots are sent to OpenAI only when you request analysis - Your API key and screenshot history remain under your control ✨ Features - GPT‑5.6 Sol, Terra, and Luna built-in models - Add and validate compatible custom vision models - Per-prompt model selection - Model-aware image detail controls, including Original for supported models - Custom system and user prompt templates - Model indicator shown beside each prompt - One-click model testing - Accurate capture on complex and GPU-heavy websites - Resizable sidebar - Fullscreen response viewer - Search, filtering, export, and screenshot history - Dark mode - No artificial usage limits—use your own OpenAI account and quota 🚀 What’s new in VisionGPT 2.0 - Migrated screenshot analysis to OpenAI’s Responses API - Added GPT‑5.6 model support - Added custom model management - Added per-prompt model testing - Added model-aware image detail selection - Removed paid license activation and validation - Improved settings, error handling, security, and screenshot workflow VisionGPT keeps your workflow local-first while giving you direct control over the OpenAI model, API key, prompts, and image-analysis settings.