GitHub
FreeAn autonomous HR agent that can answer user queries using tools
FreeFree tier
Inputs: textOutputs: text
About GitHub
An autonomous HR chatbot prototype built with ChatGPT, LangChain, Pinecone, and Streamlit. The agent can answer employee queries using three tools: a timekeeping policy document (via vector embeddings), a pandas dataframe of employee data (via Python code execution), and a calculator chain. A modernized v2 version using OpenAI's Responses API (no LangChain, no Pinecone) is also included.
Key Features
Uses LangChain agents and tools to orchestrate HR queries
Integrates Pinecone vector database for HR policy document retrieval
Streamlit frontend with chat component for user interaction
Supports both local and Azure OpenAI backends
Includes a modernized v2 version using OpenAI's Responses API (no LangChain, no Pinecone)
Provides calculator, employee data lookup, and policy query capabilities
Open-source with a free GitHub repository and video demo
Pros & Cons
Pros
- Open-source and free to use
- Modular architecture with multiple tools (policy, data, calculator)
- Supports both local and cloud (Azure) deployment
- Includes a modernized version with latest OpenAI API
- Comprehensive README with setup instructions and video demo
Cons
- Requires external API keys for OpenAI and Pinecone (free tier available)
- Prototype quality – not production-ready without further development
- Limited to HR domain and sample datasets
- Dependency on LangChain and multiple services increases complexity
Best For
Answering employee questions about HR policiesQuerying employee leave balances and supervisor informationPerforming HR-related calculations (e.g., holiday pay)Prototyping enterprise HR agent applicationsLearning how to build autonomous agents with LangChain and LLMs
FAQ
How do I install and run the autonomous HR chatbot?
Clone the repo, navigate to the directory, run 'pip install -r requirements.txt', input your API keys in hr_agent_backend_local.py, then run 'streamlit run hr_agent_frontent.py' in your terminal.
What tools does the HR agent use?
The agent uses three tools: a timekeeping policy document (embeddings stored in Pinecone), a pandas dataframe of employee data (manipulated via PythonAstREPLTool), and a calculator chain (LLMMathChain).
Is there a newer version of this project?
Yes, a modernized v2 version using OpenAI's Responses API (no LangChain, no Pinecone, with gpt-5.2 reasoning summaries) is located in the v2/ directory.