AI WORKBENCH · MULTI-MODEL WORKSPACE

odysseus-minimax

Self-hosted AI workbench extended with native MiniMax M3 text and image-01 generative capabilities.

Models

MiniMax M3 + image-01

Frontend

Modern Web UI

Bugfixes

4 Upstream Resolved

License

AGPL-3.0

Executive TL;DR — Engineering Brief

odysseus-minimax is an enhanced fork of the Odysseus workspace developed by Alex Santos. It integrates MiniMax M3 and image-01 directly into the multi-provider platform, resolving upstream routing discrepancies and UI selector bugs.

DIAGRAM & WORKFLOW

Workspace Integration Architecture

Provider auto-detection, parameter mapping, and seamless image rendering inside chat threads.

DETAILED ENGINEERING

Core Technical Pillars

Design choices, architectural patterns, and engineering decisions implemented for extreme reliability and developer experience.

Native MiniMax Provider Auto-Detection

Resolves minimaxi.com and minimax.io endpoints automatically

Introduces intelligent endpoint detection and credential validation for international and regional MiniMax accounts.

Full image-01 Model Parameter Mapping

Aspect ratio, seed, and multi-sample generation inside the workspace

Exposes native aspect ratio dropdowns and generation controls directly within the conversational interface.

Upstream Bugfixes & Route Normalization

Fixed duplicate path segments and model registry exclusions

Resolved 4 critical bugs including double /v1/v1/ prefixing and image model leakage into text-only selectors.

Unified Multi-Model Chat Experience

Switch between OpenAI, Claude, DeepSeek, and MiniMax in one tab

Combines leading AI models in a single self-hosted workspace with local chat history and custom system prompts.

CODE IN ACTION

Setup Command in Odysseus Console

Real-world usage, terminal configuration, and technical integration commands.

Terminal Example SHELL

# Inside Odysseus workspace terminal:
/setup minimax

# Enter your credentials when prompted:
# MiniMax API Key: ******************
# Group ID: ******************
# Default Model: minimax-m3

QUESTIONS & ANSWERS

Technical Frequently Asked Questions

Key clarifications regarding licensing, compatibility, deployment, and security.

Does this fork maintain compatibility with other providers?

Yes, support for OpenAI, Anthropic, Ollama, and DeepSeek remains intact.

How do I deploy this workspace?

Deploy via Docker Compose or native Python virtual environment.

Are chat conversations stored locally?

Yes, all chat history and parameters are stored in local SQLite databases.

Does it support vision inputs?

Yes, multimodal image queries are supported via compatible provider models.