andrewii23/ii23-skills
GitHubImage Generation
Generate images from text prompts using local, API-key-free, headless backends (GPT via Codex, Gemini via Antigravity).
Video
Watch and understand any video, from a clip to a full film: every shot indexed and seen once in timestamped grids (~7x cheaper than frame-by-frame), plus a transcript. Also culls footage.