hf-bot 0.2.0

A simple suite of CLI tools for Hugging Face

🤗 hf-bot

An intelligent, multi-provider agentic CLI playground for the Hugging Face Ecosystem. Instead of acting as a rigid shortcut router, hf-bot functions as an interactive co-pilot capable of executing local system operations, pulling real-time repository telemetry, checking disk capacity, and maintaining stateful multi-turn developer sessions.

---

🛑 About

1. diskspace (Preventing Storage Crashes): Stops you from downloading a model that will crash your hard drive halfway through. It checks the remote model size against your remaining local storage before you download.
2. vibecheck (Evaluating Project Activity): Checks monthly downloads, community likes, and lifecycle metrics so you can immediately see if a model is vibrant and maintained or a dormant archive.
3. peek (Inspecting Model Architecture): Avoids downloading heavy model weights just to check basic parameters. It instantly snatches and parses the remote config.json to show context windows, attention heads, and model classes.

Installation

pip install hf-bot

For local development mode and contributions:

git clone [https://github.com/kuyesu/hf-tool.git](https://github.com/kuyesu/hf-tool.git)
cd hf-bot
pip install -e .

Usage & Commands

Open hf-bot CLI

Omit string arguments or use the start target keyword to drop directly into a stateful interactive REPL environment loop:

hf-bot
# OR
hf-bot start

1. diskspace

Pass a Huggingface repository path (<repo_id>) to check its total weight footprint against your available local storage space:

hf-bot diskspace EleutherAI/gpt-j-6b

2. vibecheck

Pass a Huggingface repository ID (repo_id) to evaluate monthly usage trends, community traction, and lifecycle milestones:

hf-bot vibecheck EleutherAI/gpt-j-6b

3. peek

Pass a Huggingface model identifier (<model_id>) to fetch and parse its metadata parameters instantly:

hf-bot peek gpt2

Environment Setup

hf-bot is entirely model-agnostic and will seamlessly fall back depending on the environment variables exported in your terminal profile session.

Bash

To run via flagship cloud APIs (e.g., xAI Grok-4.3 Engine)

export XAI_API_KEY="your-grok-api-key"

To run via entirely FREE, local offline architectures (e.g., Ollama / LM Studio)

export LOCAL_MODEL_URL="http://localhost:11434/v1" export LOCAL_MODEL_NAME="llama3" # Or your chosen local tool-calling weight base

Private & Gated Repositories

If a repository requires authentication, hf-bot securely prompts for your Hugging Face token and saves it locally:

🔒 Authentication Needed
Enter your Hugging Face Access Token (input will be hidden): ············
✓ Success! Token validated and saved locally.

Uninstallation

pip uninstall hf-bot

License

MIT