Nemo Agent

Nemo Agent is your Python AI Coder!

nemo-demo.mp4

Features

• Runs blazing fast
• Generates Python project structures automatically using uv
• Writes Python code based on task descriptions
• Executes development tasks using AI-generated commands
• Utilizes the Ollama, OpenAI, Claude, or Gemini language models for intelligent code generation
• Ability to import reference documents to guide the task implementation
• Allows importing existing code projects in multiple languages to serve as a reference for the task
• Enables the importation of csv data files to populate databases or graphs
• Implements best practices in Python development automatically
• Writes and runs passing tests using pytest up to 80%+ test coverage
• Automatically fixes and styles code using pylint up to 7+/10
• Calculates and improves the complexity score using complexipy to be under 15
• Auto-formats the code with autopep8
• Shows the token count used for the responses
• Run via UV (uvx)

Coding Ability

• leetcode hards
• fastapi or flask APIs
• flask web apps
• streamlit apps
• tkinter apps
• jupyter notebook
• Note: Not all runs will be successful with all models

Install

OpenAI, Claude, or Gemini Install

Requirements

• Python 3.9 or higher
• OpenAI, Claude, or Gemini API KEY
• Mac or Linux
• No GPU requirement

Requirements Installation

• Install OpenAI, Claude, or GEMINI API KEY for zsh shell
  • echo 'export OPENAI_API_KEY="YOUR_API_KEY"' >> ~/.zshrc or
  • echo 'export ANTHROPIC_API_KEY="YOUR_API_KEY"' >> ~/.zshrc or
  • echo 'export GEMINI_API_KEY="YOUR_API_KEY"' >> ~/.zshrc
• pip install uv
• uvx nemo-agent - to run nemo-agent

OR

Ollama Install

Requirements

• Python 3.9 or higher
• Ollama running qwen2.5-coder:14b
• Linux with minimum spec of Ubuntu 24.04 with RTX 4070 or;
• Mac with minimum spec of Mac Mini M2 Pro with 16MB

Requirements Installation

• Ollama install instructions:
  • curl -fsSL https://ollama.com/install.sh | sh
  • ollama pull qwen2.5-coder:14b
• pip install uv
• uvx nemo-agent - to run nemo-agent

Usage

Providers

• ollama: uvx nemo-agent --provider ollama
• openai: uvx nemo-agent --provider openai
• claude: uvx nemo-agent --provider claude
• gemini: uvx nemo-agent --provider gemini

Import Reference Documentation Into Prompt

• Documentation files must be either: .md (Markdown) or .txt (Text) and be located in a folder
• uvx nemo-agent --docs example_folder

Import Existing Code Projects Into Prompt

• Code files must be either: .py (Python), .php (PHP), .rs (Rust), .js (JavaScript), .ts (TypeScript), .toml (TOML), .json (JSON), .rb (Ruby), or .yaml (YAML) and be located in a folder
• uvx nemo-agent --code example_folder

Import Data Into Prompt

• Data files must be .csv (CSV) and be located in a folder
• uvx nemo-agent --data example_folder

Prompting

CLI

• uvx nemo-agent "create a fizzbuzz script"

OR

File Prompt

• Prompt file must be markdown (.md) or text files (.txt)
• uvx nemo-agent --file example.md or
• uvx nemo-agent --file example.txt

Run Generated Program

• cd generated_project_folder
• source .venv/bin/activate
• python main.py

Tests

Tests are automatically created and run.

Skipping Tests

You many want to skip tests especially if you are generating a UI application.

• uvx nemo-agent "create a fizzbuzz script" --tests False

Models

Default Models

• ollama is qwen2.5-coder:14b
• openai is gpt-4o
• claude is claude-3-7-sonnet-20250219
• gemini is gemini-2.5-pro-exp-03-25

Select Models

• uvx nemo-agent "my_prompt" --provider openai --model o3-mini

Supported Models

Ollama

• Supports any 128k input token models

OpenAI

• Supports o3-mini, o1-mini, o1-preview, o1, gpt-4o, and gpt-4o-mini

Claude

• Supports claude-3-7-sonnet-20250219 and claude-3-5-sonnet-20241022

Gemini

• Supports gemini-2.5-pro-exp-03-25, gemini-2.0-flash, gemini-1.5-pro, gemini-1.5-flash

Contributing

Contributions to Nemo Agent are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Disclaimer

Nemo Agent generates code using an LLM. Every run is different as the LLM generated code is different. While it strives for accuracy and best practices, the generated code should be reviewed and tested before being used in a production environment.