mahanai 3.0.0

MahanAI: terminal AI agent for NVIDIA NIM (OpenAI-compatible), tools, streaming, /api-key, safety prompts.

MahanAI Finale

Terminal AI agent using NVIDIA NIM (integrate.api.nvidia.com) with OpenAI-compatible chat, streaming, and tools (run shell commands, read/write files, list directories).

Install

pip install mahanai-finale
mahanai

API key

1. Environment: MAHANAI_API_KEY=...
2. Project .env: MAHANAI_API_KEY=...
3. Saved config (persists): inside the app run /api-key your-key or /api-key for a hidden prompt. Keys are stored under %APPDATA%\MahanAI\config.json on Windows or ~/.config/mahanai/config.json on Linux/macOS.

Use this one for now: mai-api-91023847562910

Optional

• MAHANAI_MODEL — model id from GET https://integrate.api.nvidia.com/v1/models
• MAHANAI_STREAM=0 — disable streaming (wait for full replies)
• MAHANAI_CONFIG_DIR — override directory for config.json

Commands

• /exit or /quit — leave
• /api-key [key] — save API key (omit key for prompt)
• /api-key clear — remove saved key from disk
• /help — short help

Develop

pip install -e .
python -m mahanai

Publish to PyPI

Bump version in pyproject.toml and mahanai/__init__.py, then:

pip install build twine
python -m build
python -m twine check dist/*

Upload only the new release files (example for v1.1.0):

Windows (PowerShell) — use an API token with “Entire account” or project scope:

$env:TWINE_USERNAME = "__token__"
$env:TWINE_PASSWORD = "pypi-YOUR_TOKEN_HERE"
python -m twine upload dist/mahanai_finale-1.1.0*

macOS / Linux:

export TWINE_USERNAME=__token__
export TWINE_PASSWORD=pypi-YOUR_TOKEN_HERE
python -m twine upload dist/mahanai_finale-1.1.0*

twine cannot publish without your token; keep it out of git.