[Hidden Content]

A Python package designed to automate interactions with ChatGPT without using an API key.
It allows you to send prompts programmatically and retrieve responses, including optional web search capability if available on the ChatGPT interface.

Features

🕵️‍♂️ Undetected Browsing: Uses stealth techniques to reduce bot detection.

💬 Prompt Submission: Easily send text prompts to ChatGPT.

📥 Response Retrieval: Captures the full response generated by ChatGPT.

🌐 Optional Web Search: Attempts to activate search functionality if available.

[Hidden Content]

Still works.

[Hidden Content]

Bring your own key. Use multiple providers. Stay portable.

━━━━━━━━━━━━━━━━━━━━━━

English

MultiAI Portable is a portable multi-provider AI coding agent launcher designed for users who want a clean, flexible, and local-first way to work with different AI models.

Instead of locking you into only one provider, MultiAI Portable lets you connect different AI backends such as:

• NVIDIA NIM
• OpenRouter
• Google Gemini
• OpenAI-compatible APIs
• Anthropic / Claude API, if configured by the user
• Ollama local models

This means you can use cloud models, local models, coding models, reasoning models, and custom model IDs from one portable launcher.

━━━━━━━━━━━━━━━━━━━━━━

✨ What is MultiAI Portable for?

MultiAI Portable is mainly built for:

• AI-assisted coding
• Repository analysis
• Terminal-based AI workflows
• Local project inspection
• Portable AI setup without installing a heavy desktop app
• Using your own API keys safely

It provides a coding-agent style experience with both a terminal launcher and a local dashboard interface.

━━━━━━━━━━━━━━━━━━━━━━

🧠 Model Support

MultiAI Portable is not limited to a single model.

The NVIDIA section includes 50+ model entries in the advanced model list, plus custom model ID support. Other providers may expose more or fewer models depending on your account, API key, region, provider limits, and availability.

Recommended NVIDIA setup:

Runtime Mode: openai-compatible
Recommended Model: qwen/qwen3-coder-480b-a35b-instruct
Endpoint: [Hidden Content]

Some models may work but respond slower depending on provider load, model size, backend behavior, and runtime compatibility. For coding-agent usage, Qwen Coder is currently the recommended starting point.

━━━━━━━━━━━━━━━━━━━━━━

🔥 Main Features

• Portable Windows launcher
• Local localhost dashboard
• Multi-provider support
• NVIDIA NIM support
• OpenAI-compatible runtime mode
• Ollama / local model support
• BYOK system — Bring Your Own Key
• No bundled API keys
• Secret scan tool included
• Safe public release builder included
• Direct NVIDIA API diagnostic tool included
• Custom model ID support
• Clean local configuration
• Designed for portable usage

━━━━━━━━━━━━━━━━━━━━━━

🔐 Important Security Note

MultiAI Portable does not include any API keys.

Each user must use their own provider key.

Never share your API key.
Never upload your personal config file.
Never share the working folder after entering your key.

Do not share these files or folders:

data/ai_settings.env
data/openclaude/
data/config/
data/app_data/
engine/
node_modules/
logs

Only share or use the clean release ZIP.

━━━━━━━━━━━━━━━━━━━━━━

⚠️ Is it free or unlimited?

No fake promises here.

Some providers may offer free tiers, trial access, or developer access, but limits, rate limits, model access, availability, and quota depend on each provider.

MultiAI Portable is a launcher. It does not bypass provider limits.

━━━━━━━━━━━━━━━━━━━━━━

📦 Download

Download MultiAI Portable:
[Hidden Content]

RAR PW: decodehub.org

━━━━━━━━━━━━━━━━━━━━━━

🧩 English Tutorial

Extract the ZIP

Download and extract:

MultiAI-Portable.zip

Open the extracted folder.

Start the launcher

Run:

.\START.bat on Terminal

The launcher will prepare the portable environment automatically.

Choose your provider

Select your AI provider from the setup menu.

For NVIDIA, choose:

NVIDIA NIM - [Hidden Content]

Enter your own API key

Paste your own provider API key when asked.

For NVIDIA, the key usually starts with:

nvapi-

Recommended NVIDIA model

For best coding-agent compatibility, use:

qwen/qwen3-coder-480b-a35b-instruct

Open the dashboard

From the launcher menu, choose:

Open Dashboard

Or run:

tools\Open_Dashboard.bat

Then open this address in your browser:

http://localhost:3000

Troubleshooting

401 = invalid API key
403 = model or account access issue
404 = wrong model ID or endpoint
429 = rate limit or quota

━━━━━━━━━━━━━━━━━━━━━━

MultiAI Portable
Bring your own key. Use multiple providers. Stay portable.