Skip to content

DemoGPT vs mini-agi

DemoGPT is an open-source agent framework for ai models and llm tooling. mini-agi is an open-source agent for multi-agent frameworks.

  • mini-agi has more GitHub stars (2,933 vs 1,908).
  • DemoGPT is best for prototype a python agent that combines web search with a calculator or other custom tools.
  • mini-agi is best for prototyping a small script or game from a plain-language description.
What it is
DemoGPT is an open-source agent framework for ai models and llm tooling.
mini-agi is an open-source agent for multi-agent frameworks.
Best for
Prototype a Python agent that combines web search with a calculator or other custom tools
Prototyping a small script or game from a plain-language description
Kind
Agent
Agent
Pricing
Open source
Open source
Platforms
CLI, self-hosted, API
CLI, self-hosted, Linux, macOS, Windows
Agent type
Agent framework
Open-source agent
GitHub stars
1.9K
2.9K
Last commit
Apr 2026
Nov 2026
License
MIT
MIT
Open source
Good for
Prototype a Python agent that combines web search with a calculator or other custom tools; Ask questions over a folder of PDFs or CSV files using a local Chroma index; Generate a starter Streamlit app from a plain-language description of the desired behaviour
Prototyping a small script or game from a plain-language description; Automating a multi-step local task such as generating files or a static website; Studying how a compact autonomous agent loop is structured

What DemoGPT does

  • Prebuilt tool library

    Ships with ready-made tools for web search, weather, Wikipedia, arXiv, PubMed, YouTube, Stack Overflow, URL requests and YOLO object detection.

  • Custom tool classes

    Lets developers define their own tool by subclassing a base tool class with a name, description and run method.

  • Tool-calling and ReAct agents

    Provides agent classes that choose tools in a loop and can print their reasoning and tool calls when verbose mode is on.

  • Retrieval over local files

    BaseRAG indexes PDF, TXT, CSV and JSON files into a vector store and answers questions from them.

  • Streamlit app generation

    The original pipeline turns a written instruction into LangChain code and assembles it into a Streamlit application.

What mini-agi does

  • Minimal tool set

    The agent works with a small set of commands, including shell and Python execution, file reading, and web search.

  • Chain-of-thought planning

    The agent reasons step by step and records its thoughts before choosing each command.

  • Summarized short-term memory

    Earlier context is condensed so the agent can keep working through longer objectives.

  • Optional self-critic

    A critic pass can be enabled with an environment variable to review the agent's work, at the cost of extra API requests.

  • Pluggable web search

    Search can use DuckDuckGo by default, or Parallel Search MCP or Outcrawl when configured, or be disabled entirely.

  • User confirmation prompts

    Actions can be paused for confirmation so the user can approve or abort them before they run.

DemoGPT vs mini-agi: questions

Is DemoGPT or mini-agi cheaper?

Both are open source. Compare current plans on each official site.

What is the difference between DemoGPT and mini-agi?

DemoGPT is an open-source agent framework for ai models and llm tooling. mini-agi is an open-source agent for multi-agent frameworks. mini-agi has more GitHub stars (2,933 vs 1,908).

Keep comparing