Prompt Amplifier

Transform short prompts into detailed, structured instructions using RAG

---

What is Prompt Amplifier?

Prompt Amplifier is a Python library that transforms brief user inputs into comprehensive, well-structured prompts using Retrieval-Augmented Generation (RAG).

Instead of manually crafting detailed prompts, you can:

1. Load your domain knowledge (documents, web pages, videos)
2. Write a simple prompt ("How's the deal going?")
3. Get a detailed prompt with relevant context, structure, and instructions

---

🚀 Quick Start (5 Minutes)

Step 1: Install the Library

pip install prompt-amplifier

What this does: Installs the core Prompt Amplifier library with default dependencies (TF-IDF embedder, in-memory vector store).

---

Step 2: Set Up Your API Key

API Key Required

The expand() function requires an LLM to generate expanded prompts. You need at least one API key.

Google Gemini (Recommended - Free Tier)OpenAIAnthropic

import os
os.environ["GOOGLE_API_KEY"] = "your-key-here"

Get your free key: https://aistudio.google.com/

import os
os.environ["OPENAI_API_KEY"] = "sk-your-key-here"

Get your key: https://platform.openai.com/

import os
os.environ["ANTHROPIC_API_KEY"] = "sk-ant-your-key-here"

Get your key: https://console.anthropic.com/

---

Step 3: Create Your First Expanded Prompt

# Step 3a: Set your API key
import os
os.environ["GOOGLE_API_KEY"] = "your-google-api-key"

# Step 3b: Import the library
from prompt_amplifier import PromptForge
from prompt_amplifier.generators import GoogleGenerator

# Step 3c: Create a PromptForge instance with Google Gemini
forge = PromptForge(generator=GoogleGenerator())

# Step 3d: Add your knowledge base (domain documents)
forge.add_texts([
    "POC Health Status: Healthy means all milestones are on track with positive engagement.",
    "Warning status indicates delays or risks that need immediate attention.",
    "Critical status means major blockers that require executive escalation.",
    "Key Metrics: Winscore (0-100), Feature Fit (%), Customer Engagement Score.",
])

# Step 3e: Expand a short prompt
result = forge.expand("Check the deal health")

# Step 3f: Print the result
print(result.prompt)

What each line does:

Line	Explanation
os.environ["GOOGLE_API_KEY"]	Sets your API key for Google Gemini
from prompt_amplifier import PromptForge	Imports the main class
from prompt_amplifier.generators import GoogleGenerator	Imports Google's LLM generator
PromptForge(generator=GoogleGenerator())	Creates instance with Google Gemini
forge.add_texts([...])	Adds your domain knowledge
forge.expand("...")	Transforms short prompt into detailed one

---

Step 4: See the Output

Input (what you write):

Check the deal health

Output (what you get):

**GOAL:** Generate a comprehensive deal health assessment report.

**CONTEXT:**
Based on the POC tracking data, analyze the following aspects:
- Current health status (Healthy/Warning/Critical)
- Milestone completion percentage
- Customer engagement levels
- Risk indicators

**REQUIRED SECTIONS:**

1. **Executive Summary**
   - Overall health rating
   - Key highlights

2. **Metrics Dashboard**
   | Metric | Current Value | Status |
   |--------|--------------|--------|
   | Winscore | [0-100] | [Status] |
   | Feature Fit | [%] | [Status] |
   | Engagement | [Score] | [Status] |

3. **Risk Assessment**
   - Blockers identified
   - Mitigation strategies

4. **Recommended Actions**
   - Prioritized next steps
   - Owner assignments

**INSTRUCTIONS:**
- Use specific data points from the POC tracking system
- Highlight any warning or critical indicators
- Provide actionable recommendations

Expansion Ratio: Your 4-word input became a ~150-word structured prompt (37x expansion)!

---

🔍 Search Without API Key

You can search your knowledge base without any API key:

from prompt_amplifier import PromptForge

# No API key needed!
forge = PromptForge()

# Add your documents
forge.add_texts([
    "POC Health: Healthy means milestones on track.",
    "Winscore measures deal probability from 0-100.",
    "Feature Fit shows product-customer alignment.",
])

# Search for relevant chunks
results = forge.search("deal probability")

# Print results
for r in results.results:
    print(f"Score: {r.score:.3f}")
    print(f"Content: {r.chunk.content}")
    print("---")

Expected Output:

Score: 0.892
Content: Winscore measures deal probability from 0-100.
---
Score: 0.654
Content: POC Health: Healthy means milestones on track.
---
Score: 0.423
Content: Feature Fit shows product-customer alignment.
---

---

📁 Load Documents from Files

Instead of adding text manually, load from files:

from prompt_amplifier import PromptForge
from prompt_amplifier.generators import GoogleGenerator

forge = PromptForge(generator=GoogleGenerator())

# Load a single file
forge.load_documents("./data/manual.pdf")

# Load an entire directory
forge.load_documents("./docs/")

# Check how many documents were loaded
print(f"Loaded {forge.document_count} documents")
print(f"Created {forge.chunk_count} searchable chunks")

Supported Formats:

Format	Extension	Example
Plain Text	.txt	notes.txt
PDF	.pdf	report.pdf
Word	.docx	document.docx
Excel	.xlsx	data.xlsx
CSV	.csv	records.csv
JSON	.json	config.json

---

🎯 Complete Google Colab Example

Copy this entire code block into Google Colab:

# ============================================
# PROMPT AMPLIFIER - COMPLETE EXAMPLE
# ============================================

# Step 1: Install
!pip install prompt-amplifier google-generativeai -q
print("✅ Installation complete!")

# Step 2: Set API Key
import os
os.environ["GOOGLE_API_KEY"] = "YOUR_API_KEY_HERE"  # Get from https://aistudio.google.com/
print("✅ API key set!")

# Step 3: Import
from prompt_amplifier import PromptForge
from prompt_amplifier.generators import GoogleGenerator
print("✅ Imports complete!")

# Step 4: Create PromptForge
forge = PromptForge(generator=GoogleGenerator())
print("✅ PromptForge created!")

# Step 5: Add knowledge base
forge.add_texts([
    "Company: TechCorp is a B2B SaaS company founded in 2020.",
    "Product: AI-powered analytics platform for sales teams.",
    "Pricing: Starter $99/mo, Professional $299/mo, Enterprise custom pricing.",
    "Support: Email support for Starter, 24/7 chat for Professional, dedicated CSM for Enterprise.",
    "Trial: 14-day free trial available for all plans.",
    "Integration: Connects with Salesforce, HubSpot, and 50+ other tools.",
])
print(f"✅ Added knowledge base! ({forge.chunk_count} chunks)")

# Step 6: Test different prompts
test_prompts = [
    "pricing info",
    "how to get support",
    "product overview",
]

for prompt in test_prompts:
    print(f"\n{'='*60}")
    print(f"📝 INPUT: {prompt}")
    print(f"{'='*60}")

    result = forge.expand(prompt)

    print(f"\n📊 Expansion: {result.expansion_ratio:.1f}x")
    print(f"⏱️ Time: {result.total_time_ms:.0f}ms")
    print(f"\n📄 OUTPUT:\n{result.prompt[:800]}...")

Expected Output:

✅ Installation complete!
✅ API key set!
✅ Imports complete!
✅ PromptForge created!
✅ Added knowledge base! (6 chunks)

============================================================
📝 INPUT: pricing info
============================================================

📊 Expansion: 45.2x
⏱️ Time: 2341ms

📄 OUTPUT:
**GOAL:** Provide comprehensive pricing information for TechCorp's products.

**REQUIRED SECTIONS:**

1. **Pricing Tiers Overview**
   | Plan | Price | Key Features |
   |------|-------|--------------|
   | Starter | $99/mo | Basic features |
   | Professional | $299/mo | Advanced features, 24/7 support |
   | Enterprise | Custom | Full features, dedicated CSM |

2. **Feature Comparison**
   - What's included in each tier
   - Upgrade path recommendations
...

---

📊 Understanding the Result Object

result = forge.expand("Your prompt here")

# Access different parts of the result
print(result.prompt)              # The expanded prompt (string)
print(result.original_prompt)     # Your original input
print(result.expansion_ratio)     # How much longer (e.g., 45.2x)
print(result.retrieval_time_ms)   # Time to search documents (ms)
print(result.generation_time_ms)  # Time to generate prompt (ms)
print(result.total_time_ms)       # Total time (ms)

# Access retrieved context
print(f"Used {len(result.context_chunks)} context chunks")
for chunk in result.context_chunks:
    print(f"  - {chunk.content[:50]}...")

Example Output:

Expanded prompt: **GOAL:** Generate a comprehensive...
Original: pricing info
Expansion ratio: 45.2x
Retrieval time: 12ms
Generation time: 2329ms
Total time: 2341ms
Used 3 context chunks
  - Pricing: Starter $99/mo, Professional $299/mo...
  - Trial: 14-day free trial available for all...
  - Product: AI-powered analytics platform for...

---

🔧 Troubleshooting

Error: "API key for OpenAI is missing"

Problem: You're trying to use expand() without setting an API key.

Solution:

import os

# Option 1: Use Google (has free tier)
os.environ["GOOGLE_API_KEY"] = "your-key"
from prompt_amplifier.generators import GoogleGenerator
forge = PromptForge(generator=GoogleGenerator())

# Option 2: Use OpenAI
os.environ["OPENAI_API_KEY"] = "sk-your-key"
forge = PromptForge()  # Uses OpenAI by default

Error: "TF-IDF requires at least 2 documents"

Problem: You added only 1 document, but TF-IDF needs 2+.

Solution:

# Option 1: Add more documents
forge.add_texts(["doc 1", "doc 2", "doc 3"])

# Option 2: Use Sentence Transformers (works with 1 doc)
from prompt_amplifier.embedders import SentenceTransformerEmbedder
forge = PromptForge(embedder=SentenceTransformerEmbedder())
forge.add_texts(["single document works!"])

Error: "Source not found"

Problem: The file/directory path doesn't exist.

Solution:

# Check if path exists first
import os
path = "./my_docs/"
if os.path.exists(path):
    forge.load_documents(path)
else:
    print(f"Path not found: {path}")

---

📚 Key Features

Feature	Description
10+ Document Loaders	PDF, DOCX, Excel, CSV, TXT, JSON, Web, YouTube, Sitemap, RSS
12+ Embedders	TF-IDF, BM25, Sentence Transformers, OpenAI, Google, Cohere, Voyage, Jina, Mistral
3 Vector Stores	Memory (default), ChromaDB, FAISS
6 LLM Generators	OpenAI, Anthropic, Google, Ollama, Mistral, Together AI
Evaluation Tools	Quality metrics, embedder comparison, generator benchmarking
CLI Tool	Command-line interface for quick testing

---

🎯 What's New in v0.2.1

• ✅ Input Validation: Clear errors for empty prompts, missing files
• ✅ Structured Logging: Debug visibility with logging module
• ✅ Better Error Messages: Actionable suggestions in error text
• ✅ Fixed Generators: Separate files for Anthropic, Google
• ✅ Documentation: Clearer API key requirements

---

Next Steps

• Installation

  ---

  Install Prompt Amplifier and optional dependencies

  Installation Guide

• Quick Start

  ---

  Complete tutorial with examples

  Quick Start

• User Guide

  ---

  Deep dive into all features

  User Guide

• API Reference

  ---

  Detailed API documentation

  API Reference

---

Community

• 📦 PyPI Package
• 🐙 GitHub Repository
• 🐛 Issue Tracker
• 📄 License (Apache 2.0)

---

License

Prompt Amplifier is released under the Apache 2.0 License.

Copyright © 2025 Rajesh More and Team