YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Anvesh Agent
A lightweight, terminal-only Python coding assistant powered by Cerebras and Groq. This agent takes a natural language task and returns strictly executable code, with an option to save the output to a file.
Features
- Strict Code-Only Output: No markdown, no explanations, just raw executable code.
- Multi-Provider Support: Uses Cerebras (LLaMA-3.3-70B) as the primary model with Groq (LLaMA-3.3-70B-Versatile) as a backup.
- Fast Performance: Optimized for speed with high-throughput LLM providers.
- Terminal UI: Simple, clean CLI loop.
- Save to File: Quickly save generated code using the
save filename.pycommand. - Reliability: 25-second timeouts and automatic retry logic.
Configuration
To run this agent, you need to provide API keys for the supported providers in your environment or a .env file:
CEREBRAS_API_KEY=your_cerebras_key
GROQ_API_KEY=your_groq_key
Installation
- Clone this repository.
- Install the required dependencies:
pip install httpx python-dotenv - Run the agent:
python main.py
Usage
- Run
python main.py. - Enter your coding task when prompted (e.g., "Write a python script to scrape news headlines").
- View the generated code in your terminal.
- (Optional) Type
save script.pyto save the output to a file. - Type
exitto quit.
Model Details
- Primary: Cerebras LLaMA-3.3-70B
- Backup: Groq LLaMA-3.3-70B-Versatile
License
MIT
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support