whereismyai
CoreCoder
CoreCoder is a minimal AI coding agent engine written in Python for reading, forking, and learning how agents work.
CoreCoder is a minimal AI coding agent engine consisting of 1,309 engine lines and 2,735 total lines of pure Python. It is designed for developers who want to understand how a coding agent actually works rather than act as a daily production assistant. The engine includes a main loop, model interface, context management, eight built-in tools, sessions, plan mode, hooks, and checkpoints. It supports OpenAI-compatible APIs, DeepSeek, OmniRoute, local Ollama, LiteLLM backend, and thinking models like deepseek-reasoner. Disk mutations and shell commands require user consent before running.
CoreCoder at a glance
Key features
- Read and write files
- Execute shell commands
- Spawn sub-agents
- Compact context in three tiers
- Track tokens and costs
- Search-and-replace code editing
- Stream chain-of-thought for thinking models
Use it for
- Learn how AI coding agents work from source code
- Fork and build custom AI coding agents
- Fix bugs in code files using an LLM loop
- Run automated coding agent experiments locally
Alternatives to CoreCoder
- Kilo Code — Open-source agent with any model and no markup on inference
- OpenCode — Works with 75+ model providers, local models, and existing subscriptions
- CoreSpeed — Connects agents to apps and team memory with one endpoint.
- Nodecode — It updates itself while running and rewrites its own Lisp definitions.
- pstack — Improves AI coding agent outcomes with rigorous workflows.
- Cursor — Autocomplete, inline edits and parallel agents in one editor
- Agent Skills — Encodes senior-engineer workflows and quality gates into 24 skills.
- DeepCode — Turn natural language and research papers into production-ready code with parallel agents.
- CodeGraph — It eliminates file crawling, reducing AI agent tool calls by 88% and cutting costs.
- LangChain — Compose exactly the agent your use case needs from model, tools, prompt, and middleware.
- Qoder — It features deep codebase awareness, autonomous coding agents, and cross-platform support.
- SmallCode — Runs fully local with budget-managed context and a forgiving parser for small models.
All CoreCoder alternatives