Fabric

Fabric organizes reusable AI prompts by task for people who want repeatable ways to analyze content, write, or work with code. Its community-contributed prompts, called Patterns, provide a starting point without requiring you to write each prompt yourself.

• Task-focused prompts: Summarize papers, extract ideas from YouTube videos and podcasts, explain code, improve documentation, or draft writing from an idea.
• Local and cloud models: Use Ollama or LM Studio for local inference, or connect to providers such as OpenAI, Anthropic, Google Gemini and OpenRouter, with a different model assigned to each Pattern if needed.
• Custom Patterns: Create your own prompts or override built-in ones; prompt-library updates preserve your custom Patterns, which remain private unless you share them.
• App connections: Use the command-line interface or a self-hosted REST API, including an Ollama-compatible interface that lets compatible apps access Patterns as models.
• Audio transcription: Convert audio to text through OpenAI's speech-to-text service; this capability sends audio to an external provider.

Fabric is free, open source under the MIT license, and runs on macOS, Linux and Windows, with Docker support. The software runs on your hardware, but the selected backend determines where AI processing happens. Local backends can keep model inference on your machine; cloud backends send submitted content to their provider and require internet access and provider credentials. Its Codex and Claude Code integrations connect through the corresponding authenticated services.