AI coding glossary
Clear, quotable definitions of the terms behind modern AI-assisted development.
Terms, defined
The vocabulary of vibe coding, agentic AI, and local-first development — defined in one sentence each, then unpacked.
Vibe coding
building software by describing what you want in natural language and letting an AI model write the code, instead of typing every line yourself.
Agentic AI
artificial intelligence that pursues a goal over multiple steps — planning, using tools, and checking its work — not a single one-shot response.
AI coding agent
An AI coding agent takes a software task as a goal and completes it autonomously — reading the codebase, editing code, running tests, and proposing the change.
Multi-agent orchestration
Multi-agent orchestration coordinates several specialized AI agents working in parallel on one goal, with a planner that splits the task and routes each part.
Git worktree
an additional working copy of a repository linked to the same history, letting you work on multiple branches at once without re-cloning.
Pull request
a proposed set of code changes submitted for review before being merged — the standard checkpoint where a human approves work in git.
LLM (large language model)
an AI model trained on vast text to understand and generate human-like language — and, increasingly, to write code.
Bring your own model (BYO model)
Bring your own model (BYO model) means connecting your own AI model — via a login, API key, or local model — instead of paying for one bundled into the tool.
Local-first software
Local-first software runs on your own device and keeps your data there by default, syncing to the cloud only when you choose — not stored on a vendor's servers.
AI crew
a team of specialized AI agents working together on a software task — a Director that plans, plus a Scout, Builder, Reviewer, and Tester.
AI pair programming
coding alongside an AI assistant that suggests, writes, and discusses code with you in real time — like two-developer pair programming.
Autonomous coding agent
An autonomous coding agent completes software tasks with minimal human direction — planning, writing, testing, and submitting code on its own.
Agentic development
building software with autonomous AI agents that plan, write, test, and submit code with minimal step-by-step direction — also called agentic coding or agentic dev.
Parallel AI agents
Parallel AI agents are multiple AI coding agents working at the same time on different parts of a task, usually isolated on separate git worktrees so their changes don't collide.
Background coding agent
A background coding agent works asynchronously on a task while you do something else, then surfaces the result — typically a pull request — when it's done.
Ship with a crew of your own
Start a 7-day free trial — no credit card. Have a crew opening pull requests on your machine in minutes. You’re always the one who merges.