Hey everyone! π
Iβm excited to share Shell Sage, an AI-powered terminal assistant that brings the best of local and cloud AI to your command line. Whether you prefer fully private, offline AI or fast, cloud-based models, Shell Sage adapts to your workflow! Itβs in early development (v1.0.0), and Iβd love your feedback!
π Whatβs New?
π Run AI Locally β No API Keys, No Internet Required!
- Privacy-first, offline AI with Ollama.
- Supports Llama 3, Mistral, Phi-3, and more.
- Runs entirely on your machine for security and independence.
ollama pull llama3:8b-instruct-q4_1
shellsage config --mode local --model llama3:8b-instruct-q4_1
βοΈ Or Use Cloud APIs for More Power!
- Instant access to Groq, OpenAI, Anthropic, Fireworks.ai, OpenRouter, Deepseek.
- Switch providers effortlessly:
shellsage config --mode api --provider groq
β‘ Core Features
β
Error Diagnosis β Detect and fix risky commands.
$ rm -rf /important-folder
# Shell Sage: β Dangerous! Suggestion: rm -rf ./important-folder
β
Natural Language to Command Translation
$ shellsage ask "find large files over 1GB"
# Output: find / -type f -size +1G -exec ls -lh {} \;
β Interactive Workflows β Confirm commands before execution.
π Get Started in 2 Minutes
Clone the repo:
git clone https://github.com/dheerajcl/Terminal_assistant.git
cd Terminal_assistant
./install.sh
Option 1: Local AI (Recommended for Privacy & Offline Use)
- Install Ollama:
curl -fsSL https://ollama.com/install.sh | sh
- Pull a model & configure Shell Sage:
ollama pull llama3:8b-instruct-q4_1
shellsage config --mode local --model llama3:8b-instruct-q4_1
Option 2: API Mode (For Cloud-Based Speed & Power)
- Add your API key & configure:
shellsage config --mode api --provider openai
What Iβd Love to Know
π‘ Do you prefer local AI or cloud-based models?
π‘ What features would make this indispensable for you?
π GitHub Repo: [https://github.com/dheerajcl/Terminal_assistant]
π Web-page: [https://shellsage.vercel.app/]
Thanks for checking it out! Let me know your thoughts, and feel free to contribute. π
Top comments (0)