OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

Mini Agent

Mini Agent is an MIT-licensed Python project and CLI demo for building a single AI agent around the MiniMax M2.5 model. It uses an Anthropic-compatible API and demonstrates an execution loop with file-system and shell tools, persistent session notes, conversation-history summarization, Claude Skills, MCP tool integrations and detailed logs.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Mini Agent — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Mini Agent when you want a local CLI example to study or adapt for single-agent tasks involving a project workspace, tools and a MiniMax model API.

Who it is for

  • Developers studying or adapting a single-agent execution loop.
  • Users who want a CLI agent to work within a current or specified project workspace.
  • Builders exploring file-system, shell, MCP or Claude Skills integrations in a Python agent.

Documented uses

  • Run a CLI agent with a chosen project directory as its workspace.
  • Inspect or adapt a Python execution loop that combines model calls with file and shell tools.
  • Explore session notes and context summarization for longer tasks.
  • Extend the example with documented Claude Skills or MCP integrations.
Before you choose

Requirements & limits

The project is public on GitHub under the MIT license. The documented quick start requires uv and a MiniMax API key, obtained through a MiniMax platform account.

  • The documented configuration uses MiniMax M2.5 and a corresponding MiniMax API base; another Anthropic-compatible provider is only a likely adaptation, not demonstrated.
  • Both quick-start and development modes require uv according to the README.
  • The CLI can use the current directory or a specified workspace, relevant when the agent's file and shell tools are in scope.
  • The README describes context summarization, 15 Claude Skills, MCP tool integrations and detailed logs; its persistent-memory implementation is not detailed.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install and configure the Mini Agent CLI

Set up the quick-start CLI with a MiniMax API key and model configuration.

Before you start

  • uv installed
  • A MiniMax API key from the global or China platform
  • The corresponding API base URL

Steps

  1. Install Mini Agent from the GitHub repository with the command below.

    Documented command
    uv tool install git+https://github.com/MiniMax-AI/Mini-Agent.git

  2. Run the documented setup script for macOS/Linux or Windows PowerShell to create configuration files.

  3. Edit ~/.mini-agent/config/config.yaml and set the API key, matching API base and MiniMax-M2.5 model.

  4. Run mini-agent to use the current directory as workspace, or specify a workspace with mini-agent --workspace /path/to/your/project.

Expected outcome

A configured CLI entry point that can use the current directory or a specified directory as its workspace.

Guide limitations

  • The guide requires an API key; the repository instructions say to register and sign in on the selected MiniMax platform to create one.
  • The setup instructions document MiniMax API base URLs and MiniMax M2.5; they do not demonstrate another model provider.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
MiniMax-AI/Mini-Agent · Official homepage title · retained in its source language
How we review sourcesSuggest a correction