Fabric

1. Overview

• Fabric is an open-source framework designed to augment humans using AI. It aims to solve the integration problem of AI by organizing the fundamental units of AI—the prompts themselves—into real-world task-driven units called Patterns.
• The project emphasizes human flourishing through AI augmentation, placing people and their goals at the center of technology use.
• Fabric is available for macOS, Linux, and Windows, enabling diverse environments to leverage AI-enabled prompts across tools and interfaces.
• Visual identity and sponsorship notes are part of the project ecosystem:
• Warp sponsorship badge acknowledges support from Warp for coding with multiple AI agents.
• A logo/glyph for Fabric is included as a relevant branding element.
• A representative screenshot helps illustrate Fabric’s interface and capabilities.

1. Sponsorship and Partners

• Warp’s sponsorship highlights the collaboration between Fabric and a community focused on AI-assisted development workflows. This sponsorship underscores the project’s aim to empower developers with multi-agent AI tooling.
• The partnership signals an ecosystem where tooling for AI-assisted coding, automation, and prompt engineering can thrive in a shared space.

1. What Fabric Is and Why It Matters

• Fabric is an open-source framework for augmenting humans using AI, placing emphasis on practical, task-oriented AI work rather than abstract capabilities alone.
• The central idea is that AI’s true value comes from how smoothly we can integrate AI capabilities into daily workflows, projects, and life activities.
• The framework centers on Patterns—the organized prompts that encode real-world tasks. Patterns enable users to collect, curate, and reuse AI solutions in their favorite tools, or even directly through Fabric’s command line interface.
• For command-line aficionados, Fabric can itself be the interface, enabling a lightweight yet powerful workflow for prompt-driven AI actions.
• A key tension Fabric seeks to address is “integration” rather than “capability”: AI has the power, but the friction lies in wiring it into useful, repeatable patterns.

1. Philosophy: A Human-Centered Magnifier

• Fabric embraces a guiding philosophy: AI is not a standalone object; it magnifies human creativity and capability.
• The purpose of technology, in this view, is to help humans flourish by focusing on human problems first and then applying AI as a tool to solve them.
• The project’s philosophy emphasizes breaking problems into components, enabling AI to tackle them one piece at a time, which helps manage complexity and scale responsibly.
• An accompanying visual illustrates the concept of augmented challenges and human-centric problem solving.

1. Too Many Prompts: Organizing Prompts Into Patterns

• The Fabric ecosystem acknowledges the explosion of AI prompts across the internet and tools. The same prompts may exist in many places, with varying quality and versions.
• Fabric’s core response is to provide a robust pattern-management system that helps users collect, curate, and organize prompts into usable units across contexts.
• Patterns cover a wide range of life and work activities, including:
• Extracting insights from YouTube videos and podcasts
• Writing in a user’s own voice from a seed idea
• Summarizing opaque academic papers
• Creating perfectly matched AI art prompts for written works
• Rating and curating content quality
• Summarizing long, boring content
• Explaining code or translating documentation into usable docs
• Generating social media content from input material
• The Patterns concept scales as a core value proposition: it makes AI practical by giving people a library of reusable, interoperable prompts.

1. How to Install Fabric: Quick Start Guide

• One-Line Install (Recommended)
• Unix/Linux/macOS:
  • curl -fsSL https://raw.githubusercontent.com/danielmiessler/fabric/main/scripts/installer/install.sh | bash
• Windows PowerShell:
  • iwr -useb https://raw.githubusercontent.com/danielmiessler/fabric/main/scripts/installer/install.ps1 | iex
• Manual Binary Downloads
• Latest release binaries and SHA256 hashes are available at https://github.com/danielmiessler/fabric/releases/latest
• Package Managers
• macOS (Homebrew): brew install fabric-ai
• Arch Linux (AUR): yay -S fabric-ai
• Windows: winget install danielmiessler.Fabric
• From Source
• Requires Go. Build: go install github.com/danielmiessler/fabric/cmd/fabric@latest
• Docker
• Run pre-built images to experiment quickly, or customize with your own setup. Documentation outlines how to run the REST API server or set up Fabric in a containerized environment.
• Environment Variables and Setup
• Fabric recommends a setup step to configure directories and keys before first use.
• Visual references and live documentation are available to guide installation and use.

1. REST API Server and Ollama Compatibility

• Fabric ships with a built-in REST API server that exposes core functionality over HTTP. Start it with: fabric --serve
• Endpoints cover chat completions (including streaming), pattern management, context and session management, model/vendor listing, YouTube transcript extraction, and configuration management.
• For developers seeking Ollama compatibility, Fabric can run in Ollama-compatible mode: fabric --serve --serveOllama. This exposes endpoints such as:
• GET /api/tags to list available patterns as models
• POST /api/chat for chat completions
• GET /api/version for server version
• In Ollama-compatible setups, patterns appear as models (e.g., summarize:latest), enabling a familiar interface for clients that expect Ollama endpoints.

1. Our Approach to Prompting: Patterns, System Prompts, and Markdown

• Patterns are designed to optimize prompt readability and maintainability. Fabric prefers Markdown for pattern content to maximize readability and editability by humans and AI alike.
• Patterns typically emphasize a strong System section with explicit instructions, ensuring consistent AI behavior across usage contexts.
• A typical pattern is presented with a link to the source (e.g., a system.md file) and accompanied by a visual example of a pattern.
• This approach has proven more effective than ad hoc prompt fragments, supporting predictable outcomes and easier collaboration.

1. Intro Videos and Pattern Gallery

• Fabric provides introductory videos and a gallery of patterns to illustrate usage and capabilities:
• Intro videos include material from popular educators and content creators
• A gallery highlights the “Just Use the Patterns” approach, encouraging exploration of the /patterns directory for practical prompts
• A representative visual in the Patterns gallery demonstrates how patterns can be used to generate organized, reusable AI prompts.

1. Quick Navigation and Core Topics

• The Fabric project provides a comprehensive navigation structure, including:
• What and Why
• Updates
• Recent Major Features
• Intro Videos
• Philosophy
• Installation
• Usage
• REST API Server
• Examples
• Just Use the Patterns
• Custom Patterns
• Helper Apps
• Meta
• This structure supports users in discovering Fabric’s purposes, capabilities, and pathways to adopt Patterns in their workflows.

1. Recent Major Features: A Snapshot of What’s New

• Fabric is evolving rapidly, with features and improvements rolled out frequently. Highlights from the recent major releases include:
• v1.4.447 (April 16, 2026): Claude Opus 4.7 updates, including a 1M-token context window for Claude Opus 4.7
• v1.4.437 (March 16, 2026): OpenAI Codex Plugin support to use Codex as a backend
• v1.4.417 (Feb 21, 2026): Azure AI Gateway plugin supporting AWS Bedrock, Azure OpenAI, and Google Vertex AI via a unified gateway
• v1.4.416 (Feb 21, 2026): Azure Entra ID authentication plugin with shared utilities
• v1.4.380 (Jan 15, 2026): Microsoft 365 Copilot integration for enterprise AI grounded in Microsoft 365 data
• v1.4.378 (Jan 14, 2026): Digital Ocean GenAI support with a usage guide
• v1.4.356 (Dec 22, 2025): Complete internationalization (i18n) across 10 languages
• v1.4.350 (Dec 18, 2025): Interactive API documentation with Swagger/OpenAPI UI
• v1.4.338 / v1.4.337 (Dec 2025): Abacus and Z AI vendor support
• v1.4.334 (Nov 26, 2025): Claude Opus 4.5 update
• v1.4.331 (Nov 23, 2025): GitHub Models support
• v1.4.322 (Nov 5, 2025): Interactive HTML concept maps and Claude Sonnet 4.5
• v1.4.317 (Sep 21, 2025): Portuguese language variants and broader locale support
• v1.4.314 / v1.4.311 (Sep 2025): Expanded internationalization and Azure OpenAI migration
• v1.4.309 - v1.4.303 (Sep–Aug 2025): Expanded language support and new binary targets (Linux ARM, Windows ARM) for Raspberry Pi and Windows devices
• v1.4.294 (Aug 2025): Venice AI support (privacy-first, open source)
• v1.4.291 (Aug 2025): Speech-to-text enhancements via OpenAI
• These updates illustrate Fabric’s trajectory toward broader provider support, enterprise capabilities, internationalization, and improved developer tooling.

1. Intro Videos and Practical Usage

• Intro videos (still relevant, though some were recorded when Fabric was Python-based) remain useful for understanding core concepts and getting started.
• You can refer to videos by creators (e.g., Network Chuck, David Bombal) and other tutorials to see Fabric in action and get inspiration for practical prompts and workflows.

1. Per-Pattern Model Mapping, Aliases, and Workflow Helpers

• Per-Pattern Model Mapping: You can configure model choices per pattern using environment variables such as FABRICMODELPATTERN_NAME=vendor|model. This helps maintain per-pattern mappings consistently across environments.
• Aliases for Patterns: Fabric supports a pattern-based aliasing approach to run a pattern as a direct command, e.g., summarize instead of fabric --pattern summarize. The project provides instructions to automatically generate and load aliases across shells, with optional FABRICALIASPREFIX to standardize alias names.
• pbpaste and Clipboard Integration: The docs show how to pipe input directly from the clipboard into a pattern, a common workflow for quickly transforming content. On macOS, pbpaste is used; alternatives are described for Windows (Get-Clipboard) and Linux (xclip). For Windows, aliasing pbpaste to Get-Clipboard is suggested for parity.
• Save to Markdown: There are workflows to save outputs directly to Markdown notes or Obsidian, with an example that formats outputs as YYYY-MM-DD-title.md in a user-specified Obsidian vault.

1. Custom Patterns: Personalization, Privacy, and Safety

• Fabric supports Custom Patterns to keep personal prompts separate from built-in ones, ensuring updates don’t overwrite private work.
• Setup involves selecting a Custom Patterns directory during the Fabric setup and then using that directory to create and run personal patterns.
• Custom Patterns take precedence over built-in patterns with the same name, ensuring user intent is preserved.
• They appear in fabric --listpatterns alongside built-ins, but are not affected by updates to the built-in set.
• Privacy is emphasized: Custom patterns remain private by default unless the user chooses to share them.

1. Helper Apps and Small Tools

• Fabric ships with several helper apps that extend its capabilities:
• to_pdf: Converts LaTeX to PDF, supports piping content through Fabric to generate PDFs from AI-generated content.
• to_pdf installation: Install via Go similarly to Fabric; requires a LaTeX distribution in the system PATH.
• code2context: Generates a JSON representation of a code directory to guide AI-assisted feature creation or code edits.
• generate_changelog: Generates changelogs from Git history and PRs, with optional AI-assisted summaries. It supports SQLite caching and GitHub GraphQL for efficiency.
• pbpaste: Utility for clipboard integration on macOS; the docs provide Windows and Linux equivalents and guidance for mapping to shell profiles.
• These tools illustrate how Fabric’s ecosystem extends beyond pure prompt execution to practical, file-based and workflow-oriented tasks.

1. The Web Interface: Fabric Web App

• Fabric includes a built-in web interface as a GUI alternative to the CLI. The web app provides a different entry point for managing patterns, prompts, and configurations.
• Documentation for the web app is available in the web/README.md, with setup instructions and an overview of features.

1. Meta: People, Contributions, and Support

• Special thanks to individuals who contributed to the project, including:
• Jonathan Dunn, MVP developer and GUI lead, balancing software work with a medical career.
• Caleb Sima, pivotal in deciding to open the project to the public.
• Eugen Eisler, Frederick Ros, David Peters, Joel Parish, Joseph Thacker, Jason Haddix, Andre Guerra, and others who contributed to Go adoption, web interface, architecture, and tooling.
• Primary contributors include:
• Daniel Miessler
• Jonathan Dunn
• Scott Behrens
• Andre Guerra
• These contributors are represented with avatar images linked to their GitHub profiles.
• Acknowledgments and contributor graphs are provided by community tooling to celebrate collaboration. The project was initiated by Daniel Miessler in January 2024.
• The project’s GitHub footprint and social presence are reflected in references such as the contributor graphs, and a “Support This Project” badge signals opportunities to sponsor and sustain open-source development.

1. Supporting the Project

• The Fabric project encourages community involvement and sponsorship to sustain ongoing development, improvements, and accessibility.
• A dedicated “Sponsor This Project” badge and sponsor link are included in communications to facilitate generous support for the work and its ecosystem.

1. Visual Recap: A Rich, Image-Enabled Conceptual Map

• The Fabric documentation uses multiple images to illustrate concepts, patterns, and examples:
• A project-wide branding image to anchor identity.
• The augmented-challenges visual emphasizes how human problems become solvable through structured AI augmentation.
• A screenshot illustrating the Fabric interface and summarize pattern.
• A gallery image highlighting the pattern library and “Just Use the Patterns.”
• Pattern activity and example visuals to demonstrate how a pattern appears and behaves in practice.

1. Final Thoughts: Fabric as a Living Ecosystem

• Fabric represents a living ecosystem where human intent, practical workflows, and AI capabilities converge. The emphasis on Patterns, per-pattern model control, and custom pattern isolation supports a customizable, scalable approach to AI augmentation.
• The going-forward roadmap includes broader provider support, internationalization, enterprise features, and web-based access, ensuring Fabric remains accessible to a global audience and adaptable to varied use cases.
• The project remains open to collaboration, sponsorship, and community-driven improvements as AI-assisted work becomes more embedded in everyday professional and personal activities.

Notes on images used from the Input

• Warp sponsorship badge: https://raw.githubusercontent.com/warpdotdev/brand-assets/refs/heads/main/Github/Sponsor/Warp-Github-LG-02.png
• Fabric logo: ./docs/images/fabric-logo-gif.gif
• Fabric summary screenshot: ./docs/images/fabric-summarize.png
• Augmented challenges image: https://github.com/danielmiessler/fabric/assets/50654/31997394-85a9-40c2-879b-b347e4701f06
• Patterns gallery image: https://github.com/danielmiessler/fabric/assets/50654/9186a044-652b-4673-89f7-71cf066f32d8
• Primary contributor avatars: Daniel Miessler, Jonathan Dunn, Scott Behrens, Andre Guerra (as referenced in the input) with corresponding GitHub avatar URLs

End of description