# Doorman11991/smallcode

AI coding agent optimized for small LLMs. 87% benchmark with 4B-active model.

Repository: https://github.com/Doorman11991/smallcode
Canonical: https://ross.abutalabs.com/products/smallcode
Language: JavaScript
License: MIT
License Family: permissive
Last push: 2026-08-12T20:43:58+00:00

## Health v2 (maintenance only)
Score: 75/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 97, release rhythm 86, longevity 7
- inputs: {"age_days": 107, "days_push": 21, "days_rel": 94, "gap_med": 0, "n_releases_24m": 24}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2017, forks 153 (observed 2026-08-28T04:06:06.045661+00:00)

## What it is
SmallCode is a terminal-native AI coding agent optimized for small local LLMs (8B-35B parameters) running on consumer hardware. It compensates for small-model limitations with budget-managed context, forgiving tool-call parsing, TODO-file step decomposition, and search-and-replace patch editing, all fully local with no network required.

## Use cases
- run an AI coding agent with a local 8B model on my laptop
- code assistant that works offline without cloud API calls
- coding agent for small LLMs like Ollama or llama.cpp models
- terminal AI pair programmer for consumer hardware
- agent that handles unreliable tool calling from small models
- private coding assistant with no data leaving my machine

## When to choose
- you run local models in the 8B-35B range on consumer hardware
- privacy matters and you want zero cloud API calls
- frontier-model agents like OpenCode waste context or fail on your model's tool calls
- you want a terminal-native workflow with prebuilt binaries for any OS

## When to avoid
- you use frontier models (Claude, GPT-5) which don't need these adaptations
- your model is 4B or smaller and struggles with multi-step tool use
- you need a GUI or IDE-integrated experience

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, llm-inference, cli, developer-tools, mcp
- domain: developer-tools, large-language-models, self-hosted
- platform: cli, windows, cross-platform
- tags: coding-agent, local-llm, small-models, terminal-ui, privacy-first, ollama, llama-cpp, lm-studio, ai-agents, command-line, nodejs, linux, macos

## Member repositories
- Doorman11991/smallcode (main) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:06.045661+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:00:28.991847+00:00, confidence not recorded.
  - readme: https://github.com/Doorman11991/smallcode (fetched 2026-08-28T04:06:06.045661+00:00, sha 23c2a87f8bff)
- Data as of 2026-08-30T08:39:29.467469+00:00.