smallcode
AI coding agent optimized for small LLMs. 87% benchmark with 4B-active model.
SmallCode is a terminal-native AI coding agent specifically designed to extract useful work from small local language models (8B-35B parameters) running on consumer hardware. Unlike tools built for frontier cloud models, SmallCode compensates for the limitations of smaller models through intelligent architecture including budget-managed context, forgiving multi-format tool-call parsing, search-and-replace editing primitives, and TODO-driven task decomposition. It achieves 87% benchmark performance with 4B-active models while maintaining complete privacy through fully local operation with no required network calls.
Key Features
Use Cases
- 01Running AI-assisted coding workflows entirely offline on consumer hardware with local LLMs
- 02Developing Node.js/TypeScript backends using BoneScript declarative project compilation
- 03Debugging and refactoring codebases with hybrid semantic and exact code search
- 04Automating multi-step development tasks with decomposed planning and persistent shell sessions
- 05Privacy-sensitive development work requiring zero cloud API calls
- 06Benchmarking and evaluating local model performance on coding tasks
Related Agents
View morehermes-agent
The agent that grows with you
agency-agents
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
openinterpreter
A coding agent for open models like Kimi K3
cline
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
smallcode — FAQ
What is SmallCode?+
SmallCode is a terminal-native AI coding agent optimized for small local language models (8B-35B parameters) running on consumer hardware. It compensates for small model limitations through intelligent architecture like context budget management, forgiving tool parsing, and patch-based editing.
How do I install SmallCode?+
Install globally via npm with 'npm install -g smallcode', or use the one-line installer script for prebuilt binaries that include Node.js (bash <(curl -fsSL https://raw.githubusercontent.com/Doorman11991/smallcode/master/install.sh) on Linux/macOS). You'll also need a local LLM server like LM Studio, Ollama, or llama.cpp.
What are the prerequisites for running SmallCode?+
You need Node.js 18+ (20.x or 22.x recommended for prebuilt SQLite binaries) and a local LLM server running an OpenAI-compatible endpoint. Python 3 and Git are optional, required only for the RAG indexer. No API keys are required for basic operation; cloud escalation features are entirely opt-in.
Is SmallCode free to use?+
Yes, SmallCode is open source (MIT license) and free to use. It runs entirely locally with no required cloud API calls, so there are no usage costs beyond your own compute resources.
Which AI clients does SmallCode work with?+
SmallCode is a standalone terminal agent that connects to any OpenAI-compatible local LLM server (LM Studio, Ollama, llama.cpp) or cloud provider. It's not a plugin for other AI assistants; it operates independently as its own coding agent.
What model sizes work best with SmallCode?+
The recommended range is 8B-35B parameters. Smaller models (≤4B) struggle with multi-step tool use and lose context, while larger models (>35B) don't need SmallCode's compensating architecture and are better served by tools designed for frontier models.
How do I install smallcode?+
Open the source repository on GitHub and follow its README. smallcode is a agent — MCP Agents Market links you directly to the official repo.
Is smallcode free?+
smallcode is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.