# Augment Code — AI agent platform that automates code review, ticket-to-PR delivery, security remediation, and incident response for dev teams.

> Source: The Agents Index — https://theagentsindex.com/augment-code (structured, researched, re-verified)
> Facts last verified: 2026-08-23

Augment Code is a software development agent platform built around Cosmos, its orchestration engine for running AI coding agents at scale with context, tools, and feedback loops. Alongside a CLI and IDE integrations for VS Code and JetBrains, it serves engineering teams that want automated code review, ticket-to-PR delivery, security remediation, and incident response wired directly into GitHub, GitLab, Slack, Linear, and Jira, running in managed cloud sandboxes, on a team's own infrastructure, or a mix of both.

| Fact | Value |
| --- | --- |
| Website | https://www.augmentcode.com |
| Pricing | $100/month Business, up to 50 seats |
| API | Yes |
| Best for | Engineering leaders at organizations who want AI agents running full SDLC loops end to end, with humans at the quality gates, rather than inline completion assistants. |
| Not for | Teams that need a free tier with self-serve signup, per-prompt cost transparency, or stable per-user pricing, since Augment's only self-serve plan is $100/mo flat and usage costs vary with workload. |

## Pricing

| Tier | Price |
| --- | --- |
| Business | $100/month |
| Enterprise | Custom |

## Verdict

Augment Code centers on Cosmos, an agent platform that runs coding agents through full loops (PR review, ticket-to-PR, CVE remediation, incident response, plus test coverage, migrations, and onboarding) rather than just inline completions. Native triggers into GitHub, GitLab, Slack, Linear, and Jira, plus the option to run agents on a team's own machines, make it a fit for organizations wary of vendor lock-in. Pricing is now a single Business plan at $100/mo flat for up to 50 seats (no per-seat charge), with $100/mo of usage included and pay-as-you-go top-ups. Enterprise is sales-gated. The catch is that every performance number (66%, 2-3x, 60%+, 70%+) is vendor-published, and there is no free tier or self-serve path below $100/mo.

## How it works

1. Connect Cosmos to a trigger source such as a GitHub pull request, a Linear ticket, a Slack message, a schedule, or a webhook.
2. Cosmos spins up an agent in a sandbox with Context Engine access to the full codebase and any configured MCP tools.
3. The agent completes the loop (review, PR, fix, or incident triage) and reports back through the connected tool.

## Who it's for

- Engineering teams that want routine PR review, ticket delivery, security patching, or incident triage automated instead of done by hand.
- Developers already working in VS Code or JetBrains who want an in-editor coding agent with a large codebase-aware context window.
- Organizations that need agents to run inside their own infrastructure rather than only in a vendor's cloud.

## Strengths and weaknesses

- ✓ Cosmos orchestration ties loops across PR review, ticket-to-PR, CVE remediation, and incident response, treating agents as a complete engineering workflow rather than inline completion.
- ✓ Flat $100/mo Business plan covers up to 50 seats with $100/mo of usage included and pay-as-you-go top-ups, an unusual shape in a category dominated by per-seat pricing.
- ✓ Run agents in managed cloud, on a team's own machines, or a hybrid pool of both, which keeps them inside compliance boundaries.
- ✓ Native triggers for GitHub, GitLab, Slack, Linear, Jira, schedules, and typed webhooks, with MCP for the rest of the stack.
- ✓ SOC 2 Type II on the Business plan, with CMEK, ISO 42001, SSO, SCIM, and SIEM on Enterprise for security-sensitive buyers.
- ✗ No free tier or self-serve path below $100/mo. Enterprise requires a sales call.
- ✗ All headline performance numbers (66% faster merges, 2-3x throughput, 60%+ CVEs, 70%+ incidents) are vendor-published, with no independent benchmark found.
- ✗ Per-prompt cost visibility is aggregate-only (daily usage dashboard), with no reasoning-effort controls or per-model cost breakdown.
- ✗ Customers publicly disclosed are large enterprise (Adobe, MongoDB, Snyk, Crypto.com), so smaller teams get limited peer signal.
- ⚠ The headline performance numbers (66% faster merges, 2-3x throughput, 60%+ CVEs remediated, 70%+ incidents resolved) are Augment's own published claims. No independent benchmark corroborating them was found in the sources reviewed.
- ⚠ Pricing details conflict with older third-party trackers that still describe credit-based tiers (Indie, Developer, Standard, Max). The live pricing page only lists Business and Enterprise.
- ⚠ Top-ups after the included $100/mo make monthly cost workload-dependent and not predictable per user.

## Key features

- **Cosmos Agent Runtime** — Runs coding agents through full loops with context, tools, and feedback rather than one-shot code suggestions.
- **Context Engine** — Indexes an entire codebase in real time, giving agents up to a 200,000 token context window.
- **Trigger and Automation** — Native triggers for GitHub, GitLab, Slack, Linear, Jira, schedules, and typed webhooks start loops automatically.
- **Sandboxes and Shared File System** — Agents execute in isolated sandboxes that share a file system so multiple agents can work on the same project.
- **Flexible execution** — Loops run in Augment's managed cloud, a team's own environment, laptops, or dev VMs like Codespaces and Devcontainers.
- **Rules and Memories** — .augment/rules stores project conventions so agents follow a team's architectural preferences.

## Use cases

- **Code review** — An agent reviews every GitHub pull request automatically, a loop Augment says cuts time to merge by 66%.
- **Ticket-to-PR** — A Linear or Jira ticket triggers an agent that drafts a pull request, claimed to deliver 2-3x engineering throughput.
- **Vulnerability remediation** — An agent scans for flagged CVEs and auto-remediates over 60% of them before a developer steps in.
- **Incident response** — An alert triggers an agent that triages and resolves more than 70% of incidents before on-call is paged.

## Integrations

GitHub · GitLab · Slack · Linear · Jira · Webhooks · VS Code · JetBrains · MCP

## Sources

- https://www.augmentcode.com
- https://www.augmentcode.com/pricing
- https://www.augmentcode.com/guides
- https://checkthat.ai/brands/augment-code/pricing
- https://www.qodo.ai/blog/augment-code-vs-cursor/
- https://blog.kilo.ai/p/testing-augment-codes-new-credit