Examples

This section contains a variety of practical examples showing how to build applications with Manas. Each example demonstrates specific features of the framework and includes complete code that you can run and adapt.

Getting Started Examples

RAG Usage Learn how to use the Retrieval-Augmented Generation (RAG) capabilities to enhance LLM outputs with relevant information.
Tool Usage Build an agent that can use external tools to solve problems and answer questions by interacting with APIs and services.

Advanced Examples

Research Assistant Implement a multi-agent system that can research topics, synthesize information, and generate reports.
Knowledge Base QA Create a question-answering system over your own documents and knowledge bases.

By Feature

Agent Examples

Simple Agent - Creating and configuring basic agents
Tool-Using Agent - Building agents that use external tools
Memory-Enhanced Agent - Agents with conversation memory

Flow Examples

Two-Node Flow - Simple sequential flow with two nodes
Research Flow - Complex flow for research tasks
Feedback Flow - Flow with feedback loops for refinement

RAG Examples

Basic RAG - Simple retrieval-augmented generation
Chunking Strategies - Different document chunking approaches
Multi-Vector RAG - Advanced RAG with multiple vector indices

Running the Examples

Most examples can be run with:

# Clone the repository
git clone https://github.com/arkokoley/manas.git
cd manas

# Install dependencies
pip install -e ".[all-cpu]"  # or .[all-gpu] for GPU support

# Run an example
python examples/simple_agent.py

Be sure to set the required environment variables for API keys:

export OPENAI_API_KEY=your_api_key_here
export ANTHROPIC_API_KEY=your_api_key_here

Next Steps

After exploring these examples, check out:

Getting Started Guide for a proper introduction
Core Concepts to understand the framework architecture
API Reference for detailed documentation