About Augment Code
Overview
Cosmos is Augment Code's platform for building an always-on software delivery system that moves work from spec to verification. Teams build and tune the system in natural language, with engineers in control at the checkpoints that matter.
Agents carry the repetitive middle of the software development lifecycle while engineers hold the quality gates. Connected loops hand outcomes to each other — triage feeds authoring, authoring feeds review, and postmortems file the next tickets — so the factory improves with every run.
Key Benefits
- Every pull request gets a substantive first-pass review within minutes, and senior engineers review for intent and architecture rather than line-by-line diffs, cutting time to merge by 66%.
- The ticket-to-PR loop investigates a ticket, writes the change, and exercises it end to end for 2–3× engineering throughput.
- Vulnerability remediation moves from alert to reviewed fix without a war room, as agents map the blast radius and hand the team patches to approve, automatically remediating 60%+ of CVEs.
- Incident response finishes the investigation before the on-call engineer joins, resolving 70%+ of incidents before then.
- Experts ship for every SDLC stage — Triage, Author, Review, and Verify — and each owns its slice end to end before handing off to the next.
- Runs happen in managed cloud sandboxes, on your own machines for code that cannot leave the building, or in pools of both.
How It Works
You describe the loop you want in plain language and pick the Advisor, which asks only what it needs to know. The Advisor lays out the Experts, environment, and automation, and you approve the plan rather than the prompts. Approving the plan creates the fleet and arms the trigger, so the first ticket-to-PR run kicks off on its own.
Use Cases
- Engineering leads run code review loops so senior engineers review for intent and architecture instead of line-by-line diffs.
- Security teams remediate CVEs from alert to reviewed fix without standing up a war room.
- On-call engineers receive incident investigations with the likely cause, timeline, and context already assembled.
- Platform teams assign a ticket and merge a reviewed pull request through the ticket-to-PR loop.
- Engineering managers ramp new engineers in days with onboarding loops.
- Teams standardize on their best workflow by making it everyone's default.
Why Choose This Product
Augment Code targets whole engineering organizations rather than individual developers, running software agents at scale with the context, tools, and feedback loops they need. Quality gates are first-class: work pauses for review, resumes on reply, and a person approves consequential actions. Pricing is flat rather than per-seat, and paid plans exclude AI training on your data.
Key Features
Work Dispatcher
Scans open tickets, applies a triage rubric, and dispatches the right expert as a worker.
PR Author
Takes a task description and drives it from first commit through merge.
Code Review Fleet
Pair Review, Deep Code Review, and PR Risk Analysis work a pull request end to end and post inline findings.
Tester
Exercises a change end to end and posts results, screenshots included.
Context Engine
All plansProvides codebase understanding shared across agents and teams.
Custom Experts
Every expert is a reusable template with its own environment, capabilities, and memory.
Native Triggers
Fires loops from native triggers for GitHub, GitLab, Slack, Linear, Jira and schedules, plus typed webhooks.
Isolated Sandboxes
Runs in managed cloud sandboxes, local dev VMs, Codespaces, or your own AWS and GCP cloud.
Shared Memory
Shared memory across runs, agents, and teams means one engineer's correction sharpens everyone's runs.
Human-in-the-Loop
Work pauses for review, resumes on reply, and consequential actions are approved by a person.
Augment Code Pricing
View full pricing →- $20/mo of usage across LLM, Context Engine and compute
- Up to 50 seats included at one flat price
- Top-ups, pay as you go
- Cosmos, CLI access, MCP and native tools
- SOC 2 Type II, no AI training allowed
- $100/mo of usage across LLM, Context Engine and compute
- Up to 50 seats included at one flat price
- Top-ups, pay as you go
- Cosmos, CLI access, MCP and native tools
- SOC 2 Type II, no AI training allowed
- Custom user pricing
- Bespoke usage limits
- Volume-based annual discounts
- SSO, OIDC, and SCIM support
- CMEK and ISO 42001 compliance
- Dedicated support
| Feature | Standard $20/mo | Business $100/mo | Enterprise Custom |
|---|---|---|---|
| $20/mo of usage across LLM, Context Engine and compute | |||
| Up to 50 seats included at one flat price | |||
| Top-ups, pay as you go | |||
| Cosmos, CLI access, MCP and native tools | |||
| SOC 2 Type II, no AI training allowed | |||
| $100/mo of usage across LLM, Context Engine and compute | |||
| Custom user pricing | |||
| Bespoke usage limits | |||
| Volume-based annual discounts | |||
| SSO, OIDC, and SCIM support | |||
| CMEK and ISO 42001 compliance | |||
| Dedicated support |
Pricing extracted from the product website and may change. Check the source for current details.