Prompta
Prompta is an open-source UI client designed for communicating with ChatGPT (and GPT-4). This tool comes with functionalities such as storing all your chats loc...
Last verified:
What is Prompta?
Prompta is a keyboard-centric chat interface for interacting with large language models like ChatGPT and GPT-4, designed for fast, efficient conversations. It emphasizes mobile-friendliness, full-text search across chat history, and local storage to keep all data on the user's device without cloud dependencies. Users can sync history across devices using CRDTs for conflict-free replication, leave notes on responses for quick reference, and customize system messages to tailor AI behavior.
Key features include support for the latest models updated dynamically, compatibility with local LLMs such as Llama and Mistral, and a searchable interface that lets users find past threads instantly. Built with Svelte for reactivity, SQLite via cr-sqlite for WASM-based local database management, and Tauri for lightweight desktop apps, it avoids Electron bloat. The app runs in browsers at chat.prompta.dev or as a native desktop download for macOS and other platforms.
Prompta targets power users who prioritize keyboard navigation, developers needing quick AI assistance, and anyone frustrated with clunky web UIs for AI chats. It served as a primary tool for months before being sunsetted due to competing features in apps like Raycast. Though no longer actively developed, the web app remains running, and open-source code is available on GitHub for self-hosting or forking.
Its local-first approach ensures privacy, with no server-side processing of chats beyond model APIs. Roadmap items like context-window compression and chat history awareness highlight potential for advanced usage, though development has ceased.
Prompta pricing
Pricing model: Free
Prompta is completely free with no paid plans or tiers. All features including desktop downloads and web access require only your own LLM API keys. No subscriptions, no hidden costs—local storage and sync use no server billing.
Prompta pros
- Keyboard-centric navigation for fast input
- Full-text search across all conversations
- Syncs chat history across multiple devices
- Stores data locally for complete privacy
- Supports latest LLMs with dynamic updates
- Compatible with local models like Llama
- Customizable system messages per chat
- Leave notes on AI responses for reference
- Mobile-friendly responsive design
- Lightweight Tauri desktop builds
- CRDTs enable conflict-free replication
- SQLite WASM for reliable local DB
- Svelte-powered reactive UI
- Quick launch via web or app download
- Open-source MIT license on GitHub
- No cloud dependency for core features
- Searchable threads improve retrieval
Prompta cons
- Project officially sunsetted, no updates
- Declining usage as noted by creator
- Lacks web search integration
- No support for reasoning models
- Missing MCP protocol compatibility
- Complex multi-provider API handling
- Requires right-click to open on macOS
- Needs Rust deps for desktop builds
- No context-window compression yet
- Limited to browser or Tauri platforms
Frequently asked questions about Prompta
Why was Prompta sunsetted?
Prompta was sunsetted because Raycast added support for bringing your own API keys, solving the core problem it addressed with a more polished interface. The creator noted shifts like web search in ChatGPT and emerging reasoning models made it less competitive without major reengineering.
How does chat history sync work?
Chat history syncs across devices using SQLite compiled to WASM via cr-sqlite with CRDTs for conflict-free replicated data types, ensuring seamless merging without data loss even offline.
What models does Prompta support?
It supports chatting with the latest models like GPT-4, updated dynamically, plus local LLMs such as Llama and Mistral. Users provide their own API keys or local model setups.
Is my data stored securely?
All chat history is kept locally on your device using SQLite, with no cloud storage or server transmission beyond API calls to LLMs, prioritizing user privacy.
Can I search my past conversations?
Yes, full-text search across all previous conversations and threads is a core feature, allowing instant retrieval of chats, responses, and notes.
How do I run it on desktop?
Download the latest release from GitHub for macOS or other platforms; it's a Tauri app. On macOS, right-click and select Open first time due to signing but not notarization.
What development tools are used?
Built with Svelte for the UI, TypeScript, cr-sqlite for WASM database, and Tauri for desktop. Bun is recommended for development and building.
Can I customize the AI behavior?
Yes, customize the system message for each chat to shape LLM responses. Leave notes on specific replies like 'working code' for personal annotations.
Is the source code available?
Yes, fully open-source under MIT license on GitHub at iansinnott/prompta, with 55 releases and instructions for building and developing.
What's on the roadmap?
Planned features include context-window compression (issue #1) and chat history awareness (issue #2), though development has stopped unless revived by the community.