Github
FreeCreate LLM agents in seconds with prompts, tools, and RAG
FreeFree tier
Inputs: textOutputs: text
About Github
DemoGPT AgentHub is an open-source Python library that enables developers to create, customize, and deploy LLM agents quickly using simple prompts. It provides a comprehensive suite of built-in tools (e.g., web search, weather, Wikipedia, Python execution, Bash, Arxiv, YouTube, Stack Overflow, Wikidata, and PubMed), supports custom tool creation through a base class, and integrates retrieval-augmented generation (RAG) for enhanced knowledge retrieval. Agents can be created with the ToolCallingAgent class and OpenAI models, allowing for versatile task automation and question answering.
Key Features
Create LLM agents quickly from user prompts
11+ built-in tools: TavilySearch, Weather, Wikipedia, Bash, Python, Arxiv, YouTube, Stack Overflow, Request URL, Wikidata, PubMed
Custom tool creation via BaseTool class
ToolCallingAgent for orchestrated agent execution
Integration with retrieval-augmented generation (RAG)
Simple pip installation (pip install demogpt)
Open source with MIT license
Pros & Cons
Pros
- Simple installation and usage with pip install demogpt
- Wide variety of pre-built tools covering search, code, reference, and utilities
- Modular design allows easy extension with custom tools
- Supports both standard agent calls and RAG-enhanced workflows
- Fully open source with community contributions welcome
Cons
- Primarily demonstrated with OpenAI models; no explicit support for other LLM providers shown
- Documentation is limited to the README; may lack comprehensive guides for complex use cases
- Community size is relatively modest (1.9k stars as of writing)
- Requires familiarity with Python and LLM agent concepts to fully leverage
Best For
Rapidly prototype LLM agents for question answering and task automationBuild AI assistants with integrated web search, code execution, and data retrievalCombine RAG with tool-based agents for knowledge-rich responsesCreate custom domain-specific tools for specialized agent workflows
FAQ
How do I install DemoGPT AgentHub?
Install via pip: pip install demogpt
How do I create a custom tool?
Create a class inheriting from demogpt_agenthub.tools.BaseTool and implement the __init__ and run methods. Then pass it to an agent as part of the tools list.
What built-in tools are available?
Built-in tools include TavilySearch, Weather, Wikipedia, Bash, Python, Arxiv, YouTube Search, Stack Overflow, Request URL, WikiData, and PubMed.
Can I use RAG with agents?
Yes, DemoGPT AgentHub supports combining RAG (retrieval-augmented generation) with agents for knowledge retrieval.