nanoloop banner

nanoloop

4 devlogs
4h 4m 29s

Small Python-based agent loop for nano local models!
Designed to run with local models on a Pi 5!

Access to web search & terminal tools and runs in an agentic loop!

This project uses AI

I used Gemini to help me with writing tool call definitions and using the OpenAI API to get responses.
Qwen helped me identify issues and build reliability
AI also helped me write the GitHub releases and Action release.

Demo Repository

Loading README...

nirvaannn

Shipped this project!

Hours: 2.39
Cookies: 🍪 37
Multiplier: 12.79 cookies/hr

I added new features like an edit tool and a subagent!

I did this for the last ship but it didn’t count because of Flavortown issues so i have to ship it again to claim extra time!

Nanoloop is a small agentic loop designed to work with local models!

nirvaannn

What did I do today?

  • Added a todo list tool for the LLM so keep track of tasks
  • Updated it to support API keys (for HackClub AI and other providers)
  • Fixed tool call issues
  • Tested with cloud and local models
Attachment
0
nirvaannn

What did I do just now?

  • Added support for argparse instead of just hard-coding model names and base URLs
  • Added a new subagent/bash writer tool to reduce big model context use
  • Try to make it an agentic loop but ran into some issues to solve next time
Attachment
0
nirvaannn

What did I do today:

  • Replaced the 2 small models with 1 big model
  • Simplified the shell tool user print so it only shows the base command
Attachment
0
nirvaannn

Shipped this project!

Hours: 1.69
Cookies: 🍪 6
Multiplier: 3.02 cookies/hr

I made a little agentic loop to learn how apps like Claude Code or OpenCode work!
Except it uses local LLMs running on a Pi 5

It can browse the web using DuckDuckGo search and run Terminal commands to execute code or edit files

I learned about subagents, streaming, tool calling, and subprocess in Python!

nirvaannn

I made the app, starting with Python!

  • Tested the OpenAI SDK with llama.cpp
  • Added a loop
  • Created tools (web search with DDGS, terminal commands)
  • Added token tracking
  • Used a smaller sub-model to aid in search summarization

What did I learn along the way?

  • Subprocess and Python-based process management
  • Streaming LLM responses

🎬 Video processing — check back later!

0