Back to vendors
A

Augment Code

Also known as: Auggie CLI, Context Engine, Cosmos, Harness Cosmos

Visit site
Entry priceFrom $20 per month flat (Standard, up to 50 seats, $20 of usage included)Full pricing detail

Augment Code is an AI coding platform built around its Context Engine, now part of Harness after an October 2026 asset acquisition. Its Cosmos product runs teams of specialized agents that coordinate across the software development lifecycle with shared memory, MCP and webhook integrations, and multi agent orchestration.

Augment Code is an AI coding platform aimed at large, complex codebases and at engineering organizations. In October 2026 Harness acquired Cosmos, the Auggie CLI, the Context Engine and the Augment team, and augmentcode.com now carries the Harness name while the site keeps the Augment Code brand and the IDE extensions keep shipping. Harness sells the product as the Harness Cosmos Software Factory Agent. At its center is a proprietary Context Engine that maps a codebase by structure. It tracks what calls what, what is active and what is deprecated across repositories, libraries, documentation and history.

That engine feeds Augment's agents in the IDE, the terminal, code review and Cosmos. In VS Code and JetBrains, the agent turns a plain language task into edits and pull requests, working through plans of several steps and pausing for approval before it runs terminal commands. The Auggie CLI brings the same capability to the command line and to CI pipelines in a headless mode, and Augment's own review bot reviews pull requests with full codebase awareness. Cosmos runs teams of Experts in the cloud, reusable agents with their own environment, model and memory.

Give an Expert a task, or let a ticket, pull request, alert, Slack mention, schedule or webhook start it. It works in a fresh sandboxed VM, or on the team's own machines through self hosted daemons, and brings back a pull request, a review or a finding for a person to approve. Experts hand work to each other through pull requests, a Ticket Dispatcher sends safe tickets to PR Author on a schedule, and template Experts cover PR authoring and fixing, review, incident investigation and ticket management.

Users choose among models from Anthropic, Google, OpenAI, xAI and others, and enterprise customers can bring their own model keys. The Context Engine supplies only the slice of context a task needs. It is exposed over the Model Context Protocol and an SDK, so its codebase understanding can be plugged into other agents and clients, and the platform connects to GitHub, GitLab, Azure DevOps, Jira, Linear, Slack, Microsoft Teams, PagerDuty and Datadog.

Augment carries SOC 2 Type II and ISO/IEC 42001 certification and does not train on customer code. It offers customer managed encryption, single sign on, role based access and audit logging, with a self hosted Kubernetes option in early access and data residency on Enterprise. Pricing is flat per team with usage included rather than per seat.

Augment focuses on the problems that appear when many engineers and many agents work in the same large system, such as architectural reasoning, shared knowledge and consistency across services.

Its customer stories include Pure Storage, whose story describes 2,000 engineers working at three times their earlier velocity, and Tekion, which runs persona based agents for more than 1,300 engineers, along with Intercom, MongoDB, GoFundMe, DXC Technology, Rubrik and Webflow.

From our cloud coding agents video

Devin, Factory, Jules, OpenHands: What Each Cloud Coding Agent Does Best

Our video on the 11 cloud coding agents in the index names what each one does best. For Augment Code, that is a team of specialists that hand work to each other through pull requests.

The player starts at Augment Code's part, 8:25 in. Recorded in October 2026; the profile here is live and updates as research is verified.

Vendor details

Canonical URL

https://www.augmentcode.com/

Category

Coding agent

Funding status

Harness acquired select Augment Code assets, Cosmos, the Auggie CLI products, the Context Engine and related technology, plus the team, announced 8 October 2026; Cosmos is being renamed Harness Cosmos Software Factory Agent, and augmentcode.com now carries a Harness, Inc. copyright.

Company status

acquired

Use cases & customers

Target customers

developersenterprises

Deployment options

SaaS (hosted)Self hosted daemons on the customer's own machinesSelf hosted Kubernetes (early access)VPC and single tenant (Cosmos product page)Local stdio Context Engine MCP serverCLI (npm, Node 20+)IDE (VS Code, JetBrains)ACP clients (Zed, Neovim, Emacs)CI (GitHub Actions, Jenkins)macOS desktop (Cosmos)

Integrations

As an MCP client, Augment connects GitHub, Linear, Jira and any MCP compatible tool, and the Context Engine itself is exposed as a standalone MCP server that other agents including Claude Code, Cursor, Codex and Zed can call for codebase retrieval, in a local stdio mode or a hosted cross repo mode indexing default branches via a GitHub App. First party GitHub Actions ship for PR review and PR description, ACP support connects Zed, JetBrains, Neovim and Emacs, the CLI is published on npm as @augmentcode/auggie, and enterprise service accounts issue dedicated API tokens for CI and automation.

In practice

Your payment system spans hundreds of thousands of files across dozens of services. Augment's Context Engine maps the whole codebase semantically and pulls in context from dependencies, documentation and commit history, so its agent can reason about the impact of a change across services.

A refactor needs to stay consistent across seventeen files without breaking callers. Augment's IDE agent plans the change, applies it with full dependency awareness, and opens a pull request you can review.

You're in a regulated org where code can't train external models or leave your perimeter. Augment carries SOC 2 and ISO 42001 certification, never trains on your code, and offers VPC deployment with customer-managed keys.

Agentic Index coverage score

12.0 / 14 capabilities · 86%

Integrations & Tool Calling Full

Cosmos connects GitHub and GitHub Enterprise Server, GitLab cloud and self hosted, Azure DevOps, Jira, Linear, Slack, Microsoft Teams, ClickUp, Datadog and PagerDuty, plus any server from its MCP registry, and Experts and the Auggie CLI call external tools through MCP servers and native tools. Tool permissions allow, deny or delegate each tool call to a webhook or script policy, and a Slack mention of @Augment starts an Expert session that posts back in thread.

Sourcedocs.augmentcode.com/cosmos/config-mcp, cosmos/connectors/config-slack and cli/permissionsread 2026-10-28

Workflow Orchestration Full

Cosmos runs multiple Experts side by side in the cloud, each in its own session. An Expert launches workers as separate parallel sessions and subagents, the docs state that no Expert launches another and that the pull request is the medium Experts hand work through, and a Ticket Dispatcher runs on a schedule and dispatches PR Author for tickets that are safe to attempt. The Project Builder turns a large feature into a design, a ticket breakdown and coordinated implementation, and the Auggie CLI runs subagents in parallel.

Sourcedocs.augmentcode.com/cosmos/workers-subagents, cosmos/experts-ticket-dispatcher, cosmos/experts-templates and cli/subagentsread 2026-10-28

Knowledge Grounding & RAG Full

The Context Engine is described as a live map of the entire codebase, always current, that semantically indexes code across hundreds of thousands of files and retrieves curated context from dependencies, documentation, recent changes, commit history and issues. It is served to Augment's own agents and to other agents through an MCP server in local or hosted mode and a Context Engine SDK, and the open source Context Connectors index incrementally.

Sourceaugmentcode.com/context-engine, docs.augmentcode.com/context-services/mcp/overview and context-services/context-connectors/how-it-worksread 2026-10-28

Human Oversight & Guardrails Full

In Cosmos an Expert brings back a pull request, a review or a finding for a person to approve, and nothing merges without a human at the checkpoint.

PR Author opens a draft that stays unpublished until a person marks it ready, work that needs a specification waits for explicit human approval of the spec, the Risk Analyzer auto approves only low risk changes, and Pair Reviewer and checkpoints sit beside them.

The IDE agent pauses before terminal commands and external actions unless Agent Auto mode is on, and CLI tool permissions can allow or deny a call, put it to the user, or delegate it to a webhook or script policy.

Sourcedocs.augmentcode.com/cosmos/software-factory/ticket-to-pr, augmentcode.com/product/cloud-agents, docs/using-augment/agent and cli/permissionsread 2026-10-28

Security, Identity & Governance Full

Augment holds SOC 2 Type II attestation and ISO/IEC 42001 certification, does not train on customer code, and offers customer managed encryption keys. Enterprise adds SSO and OIDC with granular access controls and audit trails, the Cosmos page lists RBAC, and service accounts act as non human identities with their own tokens. SCIM is listed on the Enterprise plan while the Cosmos admin checklist says it will roll out soon. Augment also runs a trust center at trust.augmentcode.com.

Sourceaugmentcode.com/security, augmentcode.com/pricing, augmentcode.com/product/cosmos and docs.augmentcode.com/cosmos/guides/admin-checklistread 2026-10-28

Observability & Auditability Full

Each Cosmos session records every message, agent turn and tool call, is kept without expiry and can be reopened, shared across the organization or restricted. Automations show a Run History tab and an Event Log with the worker tree visible per session, an Analytics API reports usage, and the Cosmos page lists audit logs that feed a SIEM and replayable runs, though the docs' audit logs section is still an empty placeholder.

Sourcedocs.augmentcode.com/cosmos/sessions-overview, cosmos/manage-automations, analytics/overview and augmentcode.com/product/cosmosread 2026-10-28

Memory & State Persistence Partial

Experts capture their own memory and store it as readable Markdown in the organization's shared virtual filesystem, scoped per repository, global, channel, project or user, and load it at the start of relevant work. A noisy memory model promotes a learning only once evidence is sufficient. Augment does not say how long Expert memory lasts, when it expires or how to purge it.

Sourcedocs.augmentcode.com/cosmos/experts-memory and experts-code-review-memoryread 2026-10-28

Deployment & Data Residency Full

Cosmos sessions run in Augment hosted sandboxed VMs by default, or on a team's own machines through self hosted daemons. A self hosted Kubernetes option is in early access behind a feature flag and needs outbound access, EU tenants can set EU only inference, and Enterprise lists data residency options and multi region compute. The Cosmos product page names VPC, single tenant and on premises options, but the docs name no region list.

Sourcedocs.augmentcode.com/cosmos/environments/daemons, cosmos/guides/self-hosted-k8s, augmentcode.com/product/cosmos and augmentcode.com/pricingread 2026-10-28

Prebuilt Agents, Templates & Packs Full

Augment ships fourteen Template Experts, including PR Author, PR Fixer, Deep Reviewer, Pair Reviewer, Risk Analyzer, Verifier, Incident Investigator, Feedback Triager, Project Builder, Ticket Manager, Ticket Dispatcher and Data Analyst. Skills are defined at organization, user and project scope, and custom slash commands and published examples add reusable workflows. There is no marketplace.

Sourcedocs.augmentcode.com/cosmos/experts-templates, cosmos/config-skills and cli/custom-commands-examplesread 2026-10-28

Triggers & Channel Coverage Full

Cosmos triggers turn events into Expert sessions: GitHub issues, issue comments, pull requests, pushes, check suites and workflow runs, Linear status changes, Slack mentions, GitLab and PagerDuty events, cron schedules and custom webhooks, each filtered with JSONLogic and capped at a runs per minute limit. When a trigger fires, Cosmos opens a session in a sandboxed VM and drops the raw event payload in as the first message, running as a service account identity if configured. The Cloud Agents page says a ticket, pull request or alert can start an agent.

Sourcedocs.augmentcode.com/cosmos/config-triggers, cosmos/getting-started and augmentcode.com/product/cloud-agentsread 2026-10-28

Model Flexibility & Routing Full

The user chooses among models from Anthropic, Google, OpenAI, xAI, DeepSeek, Zhipu and Moonshot AI, with Prism routing as an option. In Cosmos the choice is made per Expert or per session, and in the Auggie CLI with /model or --model. Bring your own key is available to customers with enterprise agreements.

Sourcedocs.augmentcode.com/models/available-models, cosmos/guides/admin-checklist and augmentcode.com/product/cosmosread 2026-10-28

APIs, SDKs & MCP Extensibility Full

The Auggie SDK in TypeScript and Python builds custom integrations and agents on Auggie with typed and streaming output and custom tools, the CLI runs headless with --print and auggie cloud manages Cosmos Experts, environments and sessions with service account tokens and JSON output. The Context Engine ships an SDK with an API reference and its own MCP server, codebase-retrieval, in local stdio or hosted HTTP mode.

Sourcedocs.augmentcode.com/cli/sdk, context-services/sdk/api-reference and context-services/mcp/overviewread 2026-10-28

Testing, Debugging & Optimization Partial

A Verifier Expert runs an end to end verification pass against a running environment and reports evidence backed findings, and Cosmos offers replayable runs. Both check the customer's code and the agent's output, and Augment offers no harness to evaluate or regression test agent behavior itself.

Sourcedocs.augmentcode.com/cosmos/experts-verifier and augmentcode.com/product/cosmosread 2026-10-28

Browser & Computer Use Not documented

Experts and the CLI work through the IDE, terminal, repositories, sandboxed VMs, CI and MCP connected tools. No first party page describes the agent driving a browser or operating software through a graphical interface; the MCP page names browser automation only as a category of server a customer can add.

Sourceaugmentcode.com/product/cloud-agents, augmentcode.com/product/cosmos and docs.augmentcode.com/cosmos/config-mcpread 2026-10-28

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Funding / partnershipVerified

Augment Code agreed to sell Cosmos, the Auggie CLI, the Code Context Engine and related technology to Harness, and the team behind those products is moving to Harness to keep developing them. The company says Harness's delivery platform can bring these capabilities to customers more effectively than Augment could on its own.

Bears on: Agent capability

View source
·Observability / auditabilityPartially Verified

Augment Code released Cosmos Week 30, introducing session forking to branch conversations from completed turns and cost analytics with detailed charts for enterprise organizations. The release also expanded the Model Context Protocol (MCP) catalog with Google Workspace integrations and added Claude Fable 5 to the available model picker.

Bears on: Observability / auditability

View source
·Workflow orchestrationPartially Verified

Augment Code released Auggie CLI 0.33.0, introducing cloud triggers to enable or disable persistent workflow triggers and a worker sessions view to display child worker processes. The update also improves Model Context Protocol (MCP) reliability with proactive usability probing, automated secret scrubbing, and dynamic tool list refreshing.

Bears on: Workflow orchestration

View source
View all 6 Augment Code updates →Tracked since May 2026 · Verified from public vendor sources

Pricing

From $20 per month flat (Standard, up to 50 seats, $20 of usage included)

A flat monthly team plan with a dollar amount of usage included, plus metered LLM and compute usage beyond it.

Trial available

Included quota

Standard includes $20 and Business $100 of monthly usage across LLM, Context Engine and compute, pooled across up to 50 seats with 50 concurrent sessions. Included usage resets each billing cycle and does not roll over. Enterprise usage is custom.

What is public

Augment publishes three plans: Standard at $20 per month and Business at $100 per month, both flat for up to 50 seats with the same dollar amount of usage included, and custom Enterprise. Usage covers LLM inference at the provider's rates plus a flat 40 percent fee on LLM usage (no fee on compute) and Cosmos compute time, pooled across the team, with pay as you go top ups.

Billing mechanics

A flat monthly plan price includes an equal amount of usage, pooled across up to 50 seats with no per seat charge, and resets each cycle. Usage is metered as LLM inference at provider rates plus a 40 percent fee on LLM usage and Cosmos compute time at $0.19 an hour, and top ups are bought pay as you go once the included usage runs out. The Context Engine carries no separate price.

Cost watchouts

Heavy agent use costs more than the plan price, because LLM usage carries a 40 percent fee on top of provider model rates and Cosmos compute time is billed as well.

Variable cost rationale

Usage is metered in dollars, as LLM inference at the provider's list price plus a flat 40 percent fee, and as Cosmos compute time. Each plan includes a fixed dollar amount of usage ($20 on Standard, $100 on Business), pooled across up to 50 seats, so spend beyond the plan price depends on how much the agents run. Usage past the included amount is bought as pay as you go top ups, and top ups bought outside the plan stay valid for 12 months after purchase. Enterprise usage amounts are custom.

Additional watchouts

Cost follows usage rather than seats. Model choice and task size drive LLM spend, the 40 percent fee applies to all LLM usage, and Cosmos compute time is billed on top.

Overage / add-ons

Usage beyond the included amount is bought as pay as you go top ups, metered as LLM inference at provider rates plus a flat 40 percent fee on LLM usage, and as Cosmos compute time at $0.19 an hour in five minute increments with no fee. Top ups bought outside the plan stay valid for 12 months. Enterprise usage is custom.

Sales call required

Mixed (some tiers require a call)

Free / trial

No free tier. Harness's Cosmos page offers a free trial with no credit card, with no length or credits published.

Lowest paid plan

Standard at $20 per month flat for up to 50 seats, with $20 of usage included.

Commercial notes

Enterprise adds custom user pricing, unlimited users, SSO, OIDC and SCIM, CMEK and ISO 42001 compliance, data residency options, multi region compute, granular access controls, audit trails and dedicated support.

Key ambiguities

The plan price is a floor. Actual cost depends on LLM usage at provider rates plus the 40 percent fee, and on Cosmos compute time.

Cancellation / refund

Standard and Business are self serve monthly plans, and Enterprise is custom contracted.

Support SLA / resale

Standard and Business get community support and ticket filing, and Enterprise full SLA coverage with dedicated support. The SLA does not cover trials or beta use.

Missing data

Trial terms are not published, and per task usage is not given as a benchmark.

Agentic Index verified 2026-11-03

Alternatives to Augment Code

The closest documented capability profiles to Augment Code among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.