Augment Code

Augment Code is an AI‑powered development platform that gives engineering teams agents with deep codebase understanding. Instead of just completing lines or suggesting snippets, it continuously indexes your entire repository—code, dependencies, documentation, issue trackers, and commit history—and feeds that context into its agents. Those agents work inside VS Code, JetBrains IDEs, the terminal (via Auggie CLI), Slack, and a purpose‑built macOS desktop app called Intent. The result is AI support that behaves more like a knowledgeable senior teammate than a simple autocomplete, helping professional teams ship production‑grade code with fewer surprises.

• Context Engine: Continuously indexes all relevant sources—repositories, libraries, documentation, and history—so every agent operates with full project‑level awareness instead of guessing from the open file.
• IDE Agents (VS Code & JetBrains): Turn natural language tasks into multi‑step edits and pull requests. Agents provide task lists, handle complex workflows, and maintain persistent memories across sessions.
• Intent Workspace: A macOS desktop app that coordinates multiple AI agents around a living specification. It keeps implementation aligned with requirements inside isolated, reproducible workspaces, giving teams oversight without constant manual hand‑holding.
• Auggie CLI & Slack Integration: Terminal‑first and chat‑based agents that share the same Context Engine. They support shell‑driven workflows, quick questions, and lightweight collaboration directly from the terminal or Slack.
• Context‑Aware Code Review: A review bot that deeply understands your codebase catches subtle bugs, style mismatches, and architectural issues, then offers one‑click fixes in the IDE.
• Multi‑Surface Consistency: Whether you’re in an IDE, the terminal, Slack, or reviewing a pull request on GitHub, every agent draws from the same up‑to‑date understanding of your code, so context never drifts.
• Team‑Oriented Credit Pools & BYO Models: Usage is managed at the team level with shared credit pools, and you can bring your own AI model providers inside Intent, giving organizations control over both cost and model choice.

1. Sign up and install the extension: Create an account on augmentcode.com, then add the Augment plugin to VS Code or JetBrains. For terminal‑centric workflows, install the Auggie CLI.
2. Wait for initial indexing: The Context Engine will start indexing your repositories, dependencies, docs, and issue trackers. This one‑time process ensures agents have full project awareness from day one.
3. Open Intent (optional, macOS): If you work with spec‑driven, multi‑agent workflows, download the Intent desktop app and set up a workspace with your project’s requirements; agents will coordinate around that spec.
4. Write natural language tasks in the IDE: Use comments, chat panels, or keyboard shortcuts to describe what you want—e.g., “add a payment service following our repository pattern”—and let the agent produce edits, task lists, and a pull request.
5. Review and refine: Accept, reject, or modify the suggested changes. Use the one‑click fixes from code reviews to quickly address bugs and style issues. You can always iterate with follow‑up instructions.
6. Collaborate via Slack or CLI: Connect Slack to ask questions, get summaries, or trigger simple agent tasks without leaving the conversation. Auggie CLI lets you embed AI help directly into shell‑based CI/CD or personal scripts.

Augment Code uses a credit‑based system, where each action (a task, a review, a generation) consumes a certain number of credits. Plans are structured for both individuals and teams, with the team‑level plans supporting shared credit pools and advanced security features. All paid plans include SOC 2 Type II compliance and a strict no‑training‑on‑your‑data policy. Check the official website for the latest details, as pricing may have changed.

• Indie: $20 per month – 40,000 credits, Context Engine, MCP and native tools, SOC 2 Type II, auto top‑up credits, no AI training on your data.
• Standard: $60 per month per developer – Everything in Indie plus 130,000 credits.
• Max: $200 per month per developer – Everything in Standard plus 450,000 credits.
• Enterprise: Custom pricing – Custom credit limits, Slack integration, annual volume discounts, SSO/OIDC/SCIM, SOC 2 and security reports, dedicated support, and no AI training on your data.

• Desktop IDE plugins: Available for Visual Studio Code and JetBrains IDEs (IntelliJ, WebStorm, PyCharm, etc.), with deep integration into the editor’s native workflows.
• Auggie CLI: A terminal‑first client that runs on macOS and Linux, giving access to the same context‑powered agents for shell‑based tasks, scripts, and CI/CD pipelines.
• Intent Workspace: A dedicated macOS desktop application for spec‑driven, multi‑agent orchestration with isolated, reproducible workspaces.
• Slack integration: Chat‑based agents that share the Context Engine and can be invoked from Slack channels for lightweight collaboration and quick answers.
• GitHub code review bot: Integrates directly with pull requests to deliver context‑aware reviews, and can fire one‑click fixes back to the IDE.

• Let the Context Engine finish indexing before you start working on a new project; patience pays off with far more accurate suggestions.
• In Intent, invest time in writing a clear, up‑to‑date spec—the agents’ alignment with your requirements depends on it, and isolated workspaces only help if the spec is solid.
• Use credit pools strategically: assign most credits to developers doing heavy refactoring or new feature work, and keep an eye on the team dashboards to avoid unexpected top‑ups.
• Combine IDE agents and Intent: let IDE agents handle quick edits and refactors, while Intent manages larger feature implementations that need multi‑agent coordination and spec adherence.
• For code review, treat the bot’s comments as a starting point—always validate suggestions, especially around business logic, because the bot knows the codebase but not the business context.
• When adopting the system as a team, start with a pilot project to build familiarity with spec‑driven, multi‑step workflows before rolling out across the entire engineering org.

• Mid‑size product engineering teams shipping features daily across large, multi‑language codebases and needing AI that understands cross‑file architecture.
• Enterprise organizations in regulated industries that require strict data privacy, compliance (SOC 2, SSO/SCIM), and AI assistance that never trains on their proprietary code.
• High‑growth startups that want to keep engineering velocity high while founders focus on product strategy; Intent offloads implementation, and the credit model scales with usage.
• Senior individual developers and indie hackers looking for a powerful, context‑aware coding companion that goes far beyond inline autocomplete and offers production‑grade output.
• Developer experience teams and software consultancies that want to standardize context‑aware code review and agentic workflows across multiple projects and clients.
• Engineering managers who need visibility and control over AI usage via team‑wide credit pools and support for custom model providers.