Fabric — AI Automation Patterns & Prompt Library

Popular Patterns

extract_wisdom

The signature pattern — extracts key ideas, insights, quotes, and recommendations:

cat transcript.txt | fabric --pattern extract_wisdom

summarize

Creates a concise summary with key points:

cat article.md | fabric --pattern summarize

improve_code

Reviews and improves code quality:

cat main.py | fabric --pattern improve_code

create_keynote

Turns content into presentation slides:

cat research.md | fabric --pattern create_keynote

analyze_threat

Security threat analysis from vulnerability reports:

cat cve_report.txt | fabric --pattern analyze_threat

write_essay

Generates well-structured essays on any topic:

echo "The future of AI agents" | fabric --pattern write_essay

Building Custom Patterns

Create your own in ~/.config/fabric/patterns/my_pattern/system.md:

# IDENTITY and PURPOSE
You are an expert at [your specialty].

# STEPS
1. Read the input carefully
2. [Your specific instructions]

# OUTPUT
- [Expected format]

Then: cat input.txt | fabric --pattern my_pattern

Key Stats

• 30,000+ GitHub stars
• 100+ built-in patterns
• 500+ community contributors
• Supports 10+ LLM providers
• Written in Go — fast and portable

FAQ

Q: What is Fabric? A: Fabric is a CLI framework with 100+ curated AI prompt patterns that turn complex tasks like summarization, analysis, and code improvement into simple one-line commands.

Q: Is Fabric free? A: Yes, fully open-source under MIT license. You need your own LLM API key.

Q: Can I use Fabric with local models? A: Yes, Fabric supports Ollama and any OpenAI-compatible local inference server.