xiry.

AI · GitHub repo

Fabric

Open-source framework to apply modular AI prompt patterns via the CLI.

From the Xiry library · Updated 10 October 2026

Visit github.com ↗

Fabric is an open-source framework written in Go that organizes AI prompts into modular, task-focused "Patterns". It connects directly to your terminal pipeline, letting you pipe text, code, or video transcripts through AI models using community or custom prompt templates.

What it's for

  • Extract key insights, wisdom, and takeaways from YouTube videos or podcast transcripts
  • Summarize long academic papers, essays, or articles from the command line
  • Pipe code snippets directly into patterns to generate explanations or documentation
  • Automate markdown note generation straight into personal vaults like Obsidian

Features

  • Extensive library of curated prompt 'Patterns' for work, coding, and analysis
  • Direct terminal piping support (e.g. `cat file.txt | fabric -p summarize`)
  • Built-in YouTube transcript extraction and speech-to-text processing
  • Support for dozens of AI providers including Anthropic, OpenAI, Ollama, and Bedrock
  • Built-in REST API server with Ollama compatibility mode
  • Custom pattern creation and per-pattern model configuration

Pricing

Open source. Free and open source; requires your own API keys or local LLM instances.

Details

Type
GitHub repo
Category
AI
Topic
CLI Utilities
Open source
Yes
Platforms
Command line, macOS, Linux, Windows, API
For
Developers, technical writers, and terminal power users looking to integrate LLMs into CLI workflows.
Alternative to
Aider, llm (Simon Willison)
Tags
cli, prompts, prompt engineering, llm, automation, go

Why it was saved

  • @tom_doerr, 28 January 2025: Well-known project by Daniel Miessler providing structured AI prompt patterns for tasks like summarizing and problem-solving; CLI usage with pip install.

Related

Updated