Other tools

If your tool has no dedicated guide but accepts a custom Base URL and API Key, use the values below.

Configuration reference

MiniMax exposes two protocols. Pick whichever your tool supports — most modern tools support at least one.

OpenAI-compatible protocol

Field	Value
Provider	`OpenAI Compatible` (sometimes called `Custom` or `OpenAI-format`)
Base URL	`https://api.minimax.io/v1`
API Key	Get M Plan API key
Model ID	`MiniMax-M3.1-Flash`

Anthropic-compatible protocol

Field	Value
Provider	`Anthropic Compatible` (sometimes called `Claude` or `Custom Anthropic`)
Base URL	`https://api.minimax.io/anthropic`
API Key	Get M Plan API key
Model ID	`MiniMax-M3.1-Flash`

Protocol selection

Tool style	Recommended protocol	Typical environment variables
Claude Code-style (TUI / CLI written for Anthropic)	Anthropic-compatible	`ANTHROPIC_BASE_URL` + `ANTHROPIC_AUTH_TOKEN`
Cursor / Continue / Aider / OpenAI-format IDE plugins	OpenAI-compatible	`OPENAI_BASE_URL` + `OPENAI_API_KEY`
Tools that ask for both	Either works — Anthropic-compatible is recommended for prompt-cache benefits

Dify

Dify is an open-source LLM application development platform with workflows, RAG, agents, and observability.

Install the plugin

Open Dify Cloud, sign in, and go to Settings → Workspace → Model Provider. Find Minimax and click Install.

Configure your M Plan API key

Click Add API Key, enter your M Plan API key, set API Base to https://api.minimax.io/anthropic, and leave Group ID blank.

Start using the model

Save the settings to use MiniMax-M3.1-Flash in your workflows.

Oh My Pi

Oh My Pi is an open-source terminal coding agent built on Pi, with integrated LSP, code editing, and subagents.

Option 1: Setup wizard
Option 2: Manual setup

Run:

npx -y mmx-cli@latest agent setup

Select Oh my Pi (omp). If Oh My Pi isn’t installed, select it again when prompted to choose which tools to install. The wizard runs the official installer, checks omp --version, and adds MiniMax models to Oh My Pi’s models and defaults.Then launch Oh My Pi directly:

omp

Install Oh My Pi

Follow the official Oh My Pi installation instructions, then run omp --version to verify the installation.

Configure MiniMax models

Edit or create ~/.omp/agent/models.yml (user directory\.omp\agent\models.yml on Windows), replacing <MINIMAX_API_KEY> with your M Plan API key. If the file already contains other providers, merge minimax into the existing providers mapping.

providers:
  minimax:
    baseUrl: https://api.minimax.io/anthropic
    apiKey: "<MINIMAX_API_KEY>"
    api: anthropic-messages
    auth: apiKey
    models:
      - id: MiniMax-M3.1-Flash
        name: MiniMax-M3.1-Flash
        reasoning: true
        input: [text, image]
        contextWindow: 524288
        maxTokens: 131072
        thinking:
          mode: anthropic-adaptive
          efforts: [low, medium, high, xhigh, max]
          defaultLevel: max
          requiresEffort: true

This example uses the Anthropic-compatible protocol and sets the client context window to 512K (524288 tokens). See the 512K configuration guidance. Thinking is always enabled for MiniMax-M3.1-Flash and cannot be disabled.

Verify and launch

Run the following commands to confirm that MiniMax-M3.1-Flash appears in the model list, then launch it:

omp models minimax
omp --model minimax/MiniMax-M3.1-Flash:max

To save it as the default, enter /model in a session and assign the model to the default role. For more options, see the Oh My Pi model configuration docs.

Cherry Studio

Cherry Studio is an open-source desktop client supporting 50+ LLM providers, with built-in MCP servers and 300+ assistants.

Install

Install Cherry Studio per the official docs.

Add MiniMax provider

Open Cherry Studio → click Choose other Providers → search MiniMax, then pick MiniMax.

Enter your M Plan API key

Enter your M Plan API key (the API Host is prefilled), then click Check to verify.

Pick the model

Select MiniMax-M3.1-Flash from the model list.

Chatbox

Chatbox is a free, open-source, cross-platform AI client supporting advanced LLMs, custom API keys (BYOK), and AI agents across desktop, mobile, and web.

Install Chatbox

Install Chatbox per the official docs.

Add MiniMax provider

Open Chatbox → click Settings → click Model Provider → click Add → search MiniMax, then pick MiniMax Global.

Enter your M Plan API key

Enter your M Plan API key (the API Host is prefilled), then click Check next to API Key to verify the connection.

Pick the model

Select MiniMax-M3.1-Flash from the model list.

Xcode

Xcode is Apple’s official IDE for macOS, iOS, iPadOS, watchOS, and visionOS, now with built-in Coding Intelligence.

Install Xcode

Install Xcode 26 or later from the Mac App Store.

Add a model provider

Open Xcode → top menu Xcode → Settings → Intelligence → Add a Model Provider, choose the Internet Hosted tab, and fill in:

URL: https://api.minimax.io (bare host, no path)
API Key Header: Authorization (override the default x-api-key)
API Key: Bearer <your M Plan API key> (the word Bearer, a single space, then your sk-cp-… key — get one here)
Description: MiniMax (any value)

Enable the model

Click Add. Back in the Intelligence pane, open the new MiniMax provider and enable MiniMax-M3.1-Flash.

Start chatting

Open any project, press ⌘+0 to bring up the Coding Assistant, click the edit icon at the top-left to pick MiniMax-M3.1-Flash.

Kilo Code

Kilo Code is an open-source AI coding agent for VS Code, supporting multiple LLM providers and MCP servers.

First clear the ANTHROPIC_AUTH_TOKEN and ANTHROPIC_BASE_URL env vars — otherwise they will override the configuration.

Install the extension

Search Kilo Code in the VS Code Extensions panel and install.

Configure MiniMax provider

Open Kilo Code → Settings:

API Provider = MiniMax
MiniMax Entrypoint = api.minimax.io
MiniMax API Key = your M Plan API key
Model = MiniMax-M3.1-Flash

Click Save then Done.

Zed

Zed is an open-source, high-performance multiplayer code editor written in Rust by the Atom creators.

Install Zed

Install Zed per the official docs.

Add an LLM provider

Settings → LLM Provider → +Add Provider → choose OpenAI, fill:

API URL = https://api.minimax.io/v1
API Key = your M Plan API key
Model Name = MiniMax-M3.1-Flash

Confirm your M Plan API key

Click Save Provider. Then back in the LLM Provider list, click the new MiniMax entry, re-enter your M Plan API key and press Enter to confirm.

Pick the model

Return to the agent panel, click Select a Model at the bottom-right and pick MiniMax-M3.1-Flash.

Grok CLI

Grok CLI is an open-source terminal coding agent that connects to xAI’s Grok and OpenAI-compatible providers.

First clear the OPENAI_API_KEY and OPENAI_BASE_URL env vars.

Option 1: Setup wizard
Option 2: Manual setup

Run:

npx -y mmx-cli@latest agent setup

Select Grok CLI. If Grok CLI isn’t installed, select it again when prompted to choose which tools to install. The wizard runs the official Grok Build installer, checks grok --version, preserves unrelated settings, and adds a MiniMax model to ~/.grok/config.toml.Run grok when setup finishes.

Install Grok CLI

npm install -g grok-dev

Set environment variables and launch

export GROK_BASE_URL=https://api.minimax.io/v1
export GROK_API_KEY=sk-cp-...   # from M Plan
grok --model MiniMax-M3.1-Flash

Droid

Droid is Factory’s official terminal coding agent that integrates with IDEs and developer collaboration tools.

First clear the ANTHROPIC_AUTH_TOKEN env var, otherwise it overrides the API Key in the settings file.

Install Droid

curl -fsSL https://app.factory.ai/cli | sh        # macOS / Linux
# Windows:  irm https://app.factory.ai/cli/windows | iex

Edit config file

Add the following to customModels in ~/.factory/settings.json, replacing <MINIMAX_API_KEY> with your M Plan API key:

{
  "customModels": [{
    "displayName": "MiniMax-M3.1-Flash",
    "model": "MiniMax-M3.1-Flash",
    "baseUrl": "https://api.minimax.io/anthropic",
    "apiKey": "<MINIMAX_API_KEY>",
    "provider": "anthropic",
    "maxOutputTokens": 64000
  }]
}

Launch and pick model

Launch droid, then /model and pick MiniMax-M3.1-Flash.

Qwen Code

Qwen Code is Alibaba’s open-source terminal coding agent optimized for the Qwen model family.

Install Qwen Code

Linux / macOS:

bash -c "$(curl -fsSL https://qwen-code-assets.oss-cn-hangzhou.aliyuncs.com/installation/install-qwen.sh)"

Windows (works in both Command Prompt and PowerShell):

powershell -Command "Invoke-WebRequest 'https://qwen-code-assets.oss-cn-hangzhou.aliyuncs.com/installation/install-qwen.bat' -OutFile (Join-Path $env:TEMP 'install-qwen.bat'); & (Join-Path $env:TEMP 'install-qwen.bat')"

Launch and pick provider

Run qwen in your terminal. Pick Third-party Providers → MiniMax API Key → region International.

Enter your M Plan API key

Enter your M Plan API key (prefix sk-cp-…) and press Enter.

Confirm model IDs

The default model ID should be MiniMax-M3.1-Flash — press Enter to submit and start chatting.

Open WebUI

Open WebUI is a self-hosted, open-source AI chat platform with offline support, RAG, and multi-model runners.

Install Open WebUI

Install via Python pip (requires Python 3.11 to avoid compatibility issues):

pip install open-webui
open-webui serve

Create admin account

Open http://localhost:8080 in your browser and follow the prompt to create a local admin account.

Add an OpenAI connection

Click the avatar (top-right) → Admin Panel → top Settings tab → left Connections. In the OpenAI row, click ➕ Add Connection, fill:

URL: https://api.minimax.io/v1
Auth (Bearer mode): your M Plan API key (prefix sk-cp-…)

Pick the model

Click Save and back in the chat view, pick MiniMax-M3.1-Flash from the top model selector to start chatting.

nanobot

nanobot is an ultra-lightweight open-source personal AI agent CLI from HKUDS, with chat channels, memory, and MCP support.

Install nanobot

uv tool install nanobot-ai
# or:  pipx install nanobot-ai
# or:  pip install nanobot-ai          # may need --user or --break-system-packages on macOS

Initialize

nanobot onboard

Configure MiniMax in one shot

Replace sk-cp-... with your M Plan API key:

python3 -c '
import json, os, sys
p = os.path.expanduser("~/.nanobot/config.json")
c = json.load(open(p))
c["providers"]["minimax"]["apiKey"] = sys.argv[1]
c["providers"]["minimax"]["apiBase"] = "https://api.minimax.io/v1"
c["agents"]["defaults"]["provider"] = "minimax"
c["agents"]["defaults"]["model"] = "MiniMax-M3.1-Flash"
json.dump(c, open(p, "w"), indent=2)
' sk-cp-...

Launch

nanobot agent

OpenHands

OpenHands (formerly OpenDevin) is an open-source AI software-development agent by All-Hands-AI, with TUI, web GUI, and IDE integrations.

Install OpenHands

pipx install openhands

Configure MiniMax in one shot

Replace sk-cp-... with your M Plan API key:

pipx run --spec openhands python -c '
import sys
from openhands_cli.stores.agent_store import AgentStore
AgentStore().create_and_save_from_settings(
    llm_api_key=sys.argv[1],
    settings={"llm_model": "openai/MiniMax-M3.1-Flash",
              "llm_base_url": "https://api.minimax.io/v1"},
)
' sk-cp-...

Launch

openhands

LangChain

LangChain is the open-source framework by LangChain Inc. for building LLM-powered applications, with model adapters, retrieval, agents, and tracing.

Install

pip install langchain-openai

Use MiniMax via the OpenAI-compatible adapter

Replace sk-cp-... with your M Plan API key:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="MiniMax-M3.1-Flash",
    api_key="sk-cp-...",
    base_url="https://api.minimax.io/v1",
)
print(llm.invoke("Hello").content)

TRAE

TRAE is ByteDance’s AI-native IDE with intelligent collaboration and agent automation.

Install and launch TRAE

Download and install TRAE

Download and install TRAE from its website.

Complete initial setup

Launch TRAE. On first launch, the following screen appears:

Start setup

Click Get Started to begin setup.For detailed instructions, see the TRAE setup guide.

Configure MiniMax API

Clear OpenAI environment variables before configurationBefore configuring, ensure you clear the following OpenAI-related environment variables to avoid conflicts with MiniMax API:

OPENAI_API_KEY
OPENAI_BASE_URL

In TRAE, connect to MiniMax-M3.1-Flash with your M Plan API key.

Open settings

At the top right of the chat panel, click the Settings icon, then Models.

Select the Models tab

Switch to the Models tab.

Add a model

Click + Add Model to open the model configuration dialog.Enter the following settings:

Provider: Select MiniMax-Global
Model: Select MiniMax-M3.1-Flash
API Key: Enter your M Plan API key

Save the model

Click Add Model.

Overview

Subscription

Guides

Announcements

Configuration reference

OpenAI-compatible protocol

Anthropic-compatible protocol

Protocol selection

Dify

Oh My Pi

Cherry Studio

Chatbox

Xcode

Kilo Code

Zed

Grok CLI

Droid

Qwen Code

Open WebUI

nanobot

OpenHands

LangChain

TRAE

​Configuration reference

​OpenAI-compatible protocol

​Anthropic-compatible protocol

​Protocol selection

​Dify

​Oh My Pi

​Cherry Studio

​Chatbox

​Xcode

​Kilo Code

​Zed

​Grok CLI

​Droid

​Qwen Code

​Open WebUI

​nanobot

​OpenHands

​LangChain

​TRAE

Configuration reference

OpenAI-compatible protocol

Anthropic-compatible protocol

Protocol selection

Dify

Oh My Pi

Cherry Studio

Chatbox

Xcode

Kilo Code

Zed

Grok CLI

Droid

Qwen Code

Open WebUI

nanobot

OpenHands

LangChain

TRAE