What is Augment Code?

Augment Code is an AI native coding platform for software engineering teams. It combines coding agents, codebase intelligence, CLI and IDE workflows, code review, and agent orchestration through Cosmos.

What is the Augment Context Engine?

The Context Engine provides codebase understanding for AI agents. It retrieves relevant code and organizational context so agents can work with complex repositories without depending only on broad file searches.

What is Cosmos in Augment Code?

Cosmos is Augment Code's agent orchestration platform. It lets engineering teams build software development workflows using specialized agents, triggers, isolated environments, shared context, reusable expertise, and human approval checkpoints.

Can Augment Code work from the command line?

Yes. Auggie CLI provides access to Augment's coding agent, Context Engine, and development tools directly from the terminal.

Which IDEs does Augment Code support?

Augment provides documented integrations for Visual Studio Code and JetBrains IDEs.

Does Augment Code support AI code review?

Yes. Augment supports AI powered review of GitHub pull requests and can use codebase context to evaluate changes beyond the immediate diff.

Can Augment Code automate development workflows?

Yes. Cosmos supports workflows for coding, code review, testing, ticket implementation, incident investigation, security remediation, migrations, and recurring engineering tasks.

Does Augment Code support MCP?

Yes. Augment supports MCP and native tools, allowing agents and workflows to connect with additional tools and contextual systems.

Can Augment agents run outside Augment's cloud?

Cosmos supports managed cloud sandboxes as well as workflows using an organization's own machines and development environments.

Does Augment Code train AI models on customer data?

Augment states that AI training on customer data is not allowed under its paid plans.

Augment Code

AI native coding platform with coding agents, Context Engine, CLI, code review, and Cosmos orchestration for enterprise software engineering and large production codebases.

Visit Augment Code

About Augment Code

Augment Code is an AI native coding platform designed for enterprise software engineering. It combines coding agents with deep codebase intelligence, development tools, and automation capabilities to help engineering teams build, review, verify, and maintain software across complex production environments.

A central part of the platform is the Context Engine, which maintains an understanding of codebases and retrieves relevant information for agents as they work. Augment is designed to provide context across repositories, services, code patterns, and development history rather than requiring agents to repeatedly search large amounts of source code.

Augment can be used through supported IDEs and the Auggie CLI, while Cosmos extends the platform into agent orchestration. Cosmos can coordinate specialized software agents across workflows such as coding, pull request review, testing, ticket implementation, security remediation, migrations, incident investigation, and recurring software development automations.

The platform is designed around human controlled agentic development. Teams can introduce approval checkpoints, run agents in isolated environments, connect development systems through native integrations and MCP tools, and retain human decision making for architecture, quality, security, and consequential actions.

Key Features

Context Engine: Maintains codebase understanding and retrieves relevant context so coding agents can work across large repositories and complex software systems.
Coding Agent: Handles software development tasks from prompts and can use code context, development tools, and follow up actions while implementing changes.
Cosmos Orchestration: Coordinates specialized agents and reusable workflows across coding, reviewing, testing, verification, and other software development lifecycle activities.
Auggie CLI: Brings Augment's coding agent, Context Engine, and tools into terminal based development workflows.
AI Code Review: Reviews GitHub pull requests using broader codebase context to identify correctness, architectural, and implementation issues.
Workflow Automation: Supports triggers, schedules, reusable experts, human approval checkpoints, isolated execution environments, MCP tools, and software development automations.

Pricing

Standard: $20/month

Flat team pricing for up to 50 seats with $20 of monthly usage across LLMs, Context Engine, and compute. Includes Cosmos, coding agents, CLI access, MCP and native tools, standard compute, and up to 50 concurrent sessions.
Business: $100/month

Flat team pricing for up to 50 seats with $100 of monthly usage across LLMs, Context Engine, and compute. Includes Cosmos, coding agents, CLI access, MCP and native tools, standard compute, usage analytics, and up to 50 concurrent sessions.
Enterprise: Custom pricing

Designed for organizations requiring higher volume, advanced security, dedicated support, custom usage limits, unlimited users, custom compute, multi region compute, enterprise identity management, data residency options, audit capabilities, and additional governance controls.

Standard and Business usage can be topped up as needed. LLM inference is charged at the model provider's public API list price with a 40% service fee on LLM usage. Cosmos compute usage is also included in usage calculations.

Pricing last updated: September 16, 2026 at 12:00 AM

Use Cases

Using AI coding agents to implement features, fixes, refactoring, and other development tasks across large codebases
Automating ticket to pull request, code review, testing, migration, and security remediation workflows
Giving engineering teams shared codebase context across repositories, services, development history, and organizational knowledge
Coordinating multiple specialized software agents with human approval checkpoints across the software development lifecycle

Pros & Cons

Pros:

Context Engine is designed to provide agents with relevant understanding across large and complex production codebases
Combines IDE workflows, terminal based agents, code review, agent orchestration, MCP tools, and development automation in one platform
Supports human approval checkpoints, isolated execution environments, organizational knowledge, and enterprise governance

Cons:

Advanced identity management, data residency, custom compute, and several governance capabilities require the Enterprise plan
Multi region compute and unlimited concurrent agent sessions are reserved for Enterprise deployments