A local-first vault for your AI prompt collection

What it does

AI Gist is a desktop and mobile application for collecting and deploying AI prompts. It stores everything locally by default, letting you organize prompts with tags, categories, and ratings, then invoke them with Jinja-style variable substitution. The app also enlists AI models—local or remote—to generate new prompts, rewrite existing ones, and automatically extract template variables.

The interesting bit

The project treats prompts as structured assets rather than disposable text snippets. By combining Jinja templating with automatic variable extraction and usage history, it turns a simple text collection into a reusable prompt toolkit that defaults to local storage. The cross-platform reach—Electron for desktop, Capacitor for mobile—means the library follows you without requiring a hosted backend.

Key highlights

• Local-first storage with optional WebDAV or iCloud backup; data never needs to touch a remote server by default.
• Jinja template support with dynamic variable filling and AI-powered variable extraction.
• Built-in AI integration for Ollama, LM Studio, OpenAI, and other models to generate, polish, and refine prompts.
• Multi-platform coverage across Windows, macOS, Linux, Android, and iOS.
• Full data portability via CSV export and import; AGPL-licensed and open source.

Caveats

• macOS builds currently trigger Gatekeeper security warnings that require manual intervention to launch.
• The project is primarily maintained by a single developer, so long-term velocity depends on solo bandwidth.

Verdict

Worth a look if you maintain a large personal prompt library and want it templated, searchable, and offline by default. Probably overkill if you just need a scratchpad for occasional ChatGPT queries.