# Augment Cosmos

Coding agent platform for specialized tasks and automations. Cosmos combines Experts, repositories and tools in local or managed environments.

Canonical profile: https://agenten-kompass.mmeierhoff.chatgpt.site/en/agents/augment-cosmos/
Editor: [Michael Meierhoff](https://agenten-kompass.mmeierhoff.chatgpt.site/en/about/)
Sources reviewed: 2026-10-09
Basis: provider documentation. Examples are editorial suggestions, not hands-on product tests.

## Overview

Cosmos organizes specialized Experts and agent sessions. Auggie CLI and Augment offerings for VS Code and JetBrains belong to the same product family; this profile focuses on Cosmos.

- Provider: Augment Code
- Product family: Augment Cosmos
- Profile type: Coding agent platform with Experts and automations
- Environment: Cosmos web surface and connected services; Auggie CLI and IDE offerings as related interfaces
- Connections: GitHub, Slack, Linear, PagerDuty, webhooks, schedules, MCP and skills
- Setup: Configure the organization, GitHub app, repository access and agent environment
- Pricing model: Standard USD 20/team/month including USD 20 usage; Business USD 100 including USD 100, each for up to 50 seats

## Operation & oversight

Operational details for this profile have not yet been researched.

## Provider

Augment Computing, Inc. operates Augment Code for agent-assisted software development. Its offering includes Cosmos, Auggie CLI, IDE integrations and a shared Context Engine for project knowledge.

- Operator: Augment Computing, Inc.
- Organization: Company
- Website: https://www.augmentcode.com
- Reviewed: 2026-10-09

- Provider source: https://www.augmentcode.com/legal/privacy-policy

- Provider source: https://docs.augmentcode.com/introduction

## Prices and included usage

Pricing checked: 2026-10-09

A flat monthly team price includes a usage budget for models, Context Engine and compute.

### Standard

USD 20 per team per month

Up to 50 seats and USD 20 included monthly usage budget.

Price source: https://www.augmentcode.com/pricing

### Business

USD 100 per team per month

Up to 50 seats and USD 100 included monthly usage budget.

Price source: https://www.augmentcode.com/pricing

### Enterprise

Custom pricing

Larger teams and contractual enterprise requirements.

Price source: https://www.augmentcode.com/pricing

LLM inference is billed at the model provider’s public API list price. Context Engine and compute also consume the budget; top-ups outside the plan remain valid for twelve months.

Pricing details: https://www.augmentcode.com/pricing

## Key features

### Specialize Experts

Experts combine reusable instructions, model selection, tools and an environment for specific tasks.

Source: https://docs.augmentcode.com/cosmos/getting-started

### Connect repositories in environments

An environment can contain multiple repositories and the required development configuration.

Source: https://docs.augmentcode.com/cosmos/configuring-cosmos

### Use cloud or your own host

Agents can work in managed virtual machines or on a self-configured laptop, server or VM.

Source: https://docs.augmentcode.com/cosmos/configuring-cosmos

### Use events and schedules

GitHub, Slack, Linear or PagerDuty events, webhooks and cron schedules can trigger agent work.

Source: https://docs.augmentcode.com/cosmos/getting-started

### Add capabilities and skills

Capabilities provide tools and MCP access; skills add reusable task instructions.

Source: https://docs.augmentcode.com/cosmos/getting-started

### Private and shared files

Virtual files can remain private or be shared within the organization and extend experts’ context.

Source: https://docs.augmentcode.com/cosmos/getting-started

### Advisor for setup

Advisor supports configuring experts, integrations, environments and automations.

Source: https://docs.augmentcode.com/cosmos/getting-started

## Requirements and limits

### Configure the GitHub app administratively

Requirement: Repository access is configured through the installed GitHub app and the repositories selected by an administrator.

Source: https://docs.augmentcode.com/cosmos/configuring-cosmos

### Assign integration permissions explicitly

Requirement: Slack and Linear connections require suitable installation and permissions. Authorized channels and events determine access.

Source: https://docs.augmentcode.com/cosmos/configuring-cosmos

### Prepare the agent environment

Requirement: Repositories, the base image or host, and required tools must be configured before the agent run.

Source: https://docs.augmentcode.com/cosmos/configuring-cosmos

### Standard and Business up to 50 seats

Limitation: Standard and Business each include up to 50 seats. The included budget covers usage; LLM and Context Engine consumption are billed.

Source: https://www.augmentcode.com/pricing

## Example workflows

### Review a pull request with an Expert

- Your input: An accessible repository, a pull request and an Expert with clear review rules.
- Possible result: An agent session with reasoned findings and references.
- What to check: Check the PR state, tools, references and effects of any comments.

### Use a ticket event as a coding task

- Your input: An authorized Linear connection, a bounded trigger and a configured project environment.
- Possible result: A triggered Expert task with a traceable work history.
- What to check: Check event filters, repository permissions, changes and results.

## Context

Repository access is configured through the installed GitHub app and the repositories selected by an administrator. Slack and Linear connections require suitable installation and permissions. Authorized channels and events determine access.

Official product documentation reviewed. Workflows and assessments are editorial; hands-on product tests are outside the scope of this directory.

## Sources

- https://docs.augmentcode.com/introduction

- https://docs.augmentcode.com/cosmos/getting-started

- https://docs.augmentcode.com/cosmos/configuring-cosmos

- https://www.augmentcode.com/pricing

[Sources and method](https://agenten-kompass.mmeierhoff.chatgpt.site/en/sources/)
