
Augment Code
Coding agents and a Context Engine for large codebases
About Augment Code
Augment Code is an AI coding platform for engineering teams that work in large, production codebases. Its centre is Cosmos, an agent orchestration platform for what Augment calls a software factory: connected loops in which agents carry the repetitive middle of the development lifecycle, such as turning tickets into pull requests, reviewing every pull request, preparing fixes for vulnerability alerts and investigating incidents, while engineers keep the quality gates. Alongside Cosmos sit the Auggie CLI for the terminal, extensions for VS Code and JetBrains IDEs, and the Context Engine they all share.
You describe the loop you want to the Cosmos Advisor in plain language. It asks only what it needs, then proposes a plan: which Experts do the work, which environment they run in and what triggers them. You approve the plan rather than writing prompts, and the loop goes live.
Each Expert is a reusable template with its own instructions, model, tools, environment and memory.
- •Cosmos Experts and Loops - Experts hand work to each other, so triage feeds authoring, authoring feeds review and postmortems file the next tickets.
- •Context Engine - Semantic understanding of your repositories that powers every Augment agent, also offered to other tools through an MCP server, with quickstarts for Claude Code, Codex, Cursor and others, and an experimental SDK.
- •Auggie CLI - Installs from npm on macOS, Linux and Windows WSL.
- •IDE Extensions - VS Code and JetBrains extensions bring agent mode, chat, a task list and completions into the editor; a Vim and Neovim plugin adds chat and completions.
- •Code Review - AI review of GitHub pull requests on all plans; the enterprise version adds analytics, user allowlists, MCP context from Jira, Linear and Notion, and unlimited seats and repositories.
Augment Code is aimed at engineering organizations with large codebases that want agents doing real delivery work. Platform teams can automate review on every pull request, turn well-specified tickets into draft pull requests, work through vulnerability backlogs and give on-call engineers a head start on incidents, with human approval where it matters.
Pricing
$20 - $100/mo
- Standard$20/mo
- Business$100/mo
- EnterpriseContact sales













































