What Atlassian's Rovo Dev CLI Does (and Doesn't)

atlassian-cli is an independent, community open-source project, not affiliated with, endorsed by, or sponsored by Atlassian, and is not Atlassian's official CLI (acli). Product names are used only to identify compatibility.

What is the Rovo Dev CLI?

The Rovo Dev CLI is Atlassian's official agentic AI coding agent that runs inside your terminal. It is shipped as an extension to Atlassian's official command-line tool, acli, and is invoked with acli rovodev. Think of it as a terminal chat-and-act loop: you describe a task in natural language, and the agent reads your codebase, proposes changes, runs commands, and can reach into your Atlassian workspace to act on Jira work items or Confluence pages.

Rovo Dev CLI is one surface of Rovo Dev, Atlassian's wider push into AI for software teams. It sits in the same category as other terminal coding agents that became popular through 2025, and Atlassian has publicly highlighted its performance on open coding benchmarks such as SWE-bench. The distinguishing pitch is enterprise integration: because it comes from Atlassian, it is built to plug directly into Jira, Confluence, and the rest of the Atlassian toolchain.

What the Rovo Dev CLI does

Once installed and authenticated, the Rovo Dev CLI behaves like an interactive agent. Based on Atlassian's own documentation and announcements, its headline capabilities include:

• Codebase understanding. Ask it to explain an unfamiliar repository, trace how a feature works, or summarize what a module does.
• Feature implementation. Describe a change and it drafts the edits across files, then lets you review before applying.
• Migrations and refactors. It is positioned for the tedious, repetitive kind of change that spans many files.
• Atlassian context. Because it is wired into your Atlassian site, you can act on Jira work items and update Confluence documentation without leaving the shell.
• MCP and web research. It can connect to Model Context Protocol servers to pull in outside context during a task.

Getting started follows the pattern Atlassian documents. The Rovo Dev CLI is an extension for acli, so you install or update acli first, authenticate, then launch an interactive session:

# Rovo Dev CLI ships inside acli (Atlassian's official CLI).
# 1. Install or update acli using Atlassian's instructions for your OS.

# 2. Authenticate the Rovo Dev agent to your Atlassian site.
acli rovodev auth login

# 3. Start an interactive agent session in your project directory.
acli rovodev run

From there it is a conversation. You type a request, the agent works, and you approve or reject its proposed actions. That interaction model, natural-language in, reviewed changes out, is the core of what makes it an "agent" rather than a command you script.

What the Rovo Dev CLI doesn't do

The "doesn't" side matters just as much, because it is where people's expectations most often break.

It is not a deterministic API client. The Rovo Dev CLI is a probabilistic AI agent. Give it the same prompt twice and you may get two different sets of edits. That is fine for exploratory coding, but it is the wrong tool when you need the exact same result every run, such as a CI gate, a nightly export, or a bulk field update that must touch a precise set of issues.

It is not free anymore. The Rovo Dev CLI was free to use during its beta. Atlassian's current documentation states there is no free tier: using it requires a paid Rovo Dev subscription, which allots a monthly pool of Rovo Dev credits per developer. Credits are consumed as you run tasks, so heavy use has an ongoing cost. Always check Atlassian's own pricing page for the current terms, because they change.

It is not a thin, single-purpose scripting binary. The value of an AI agent is judgment and flexibility, not predictability. If your job is "transition every issue matching this JQL to Done, exactly, every time," you do not want an agent reasoning about it. You want a plain command that does precisely that and nothing else.

It is not this project. To be explicit: atlassian-cli, the tool this site documents, is not the Rovo Dev CLI and is not affiliated with Atlassian. The two are complementary, not competing.

Rovo Dev CLI vs a scriptable API client

The clearest way to understand the Rovo Dev CLI is to place it next to a deterministic command-line client. The table below compares the Rovo Dev CLI with atlassian-cli on factual, checkable axes (type, vendor, delivery, license, and primary purpose), not invented performance numbers.

Neither column is "better." They are answers to different questions. One brings AI judgment to your code; the other brings repeatable, auditable automation to your Atlassian data. For a broader look at the deterministic-client side, see the atlassian-cli guide and the sibling explainer on acli vs atlassian-cli.

When to reach for which

Scripting Atlassian without an AI agent

If what you actually need is deterministic automation rather than an AI agent, you do not need the Rovo Dev CLI at all. A plain API client covers it. Here is what that looks like with atlassian-cli, the independent binary this site documents. Every command below produces the same result every time, which is exactly what an agent cannot promise.

# Search Jira with JQL, pipe machine-readable JSON to jq
atlassian-cli jira issue search \
  --jql "project = DEV AND status = Open" \
  --format json | jq '.[].key'

# Bulk transition, always preview with --dry-run first
atlassian-cli jira bulk transition \
  --jql "project = DEV AND status = 'In Progress'" \
  --transition "Done" \
  --dry-run

# Export a Confluence space to a JSON backup
atlassian-cli confluence bulk export \
  --cql "space = DEV" \
  --output backup.json --format json

# List open Bitbucket pull requests for a repo
atlassian-cli bitbucket --workspace myteam pr list api-service \
  --state OPEN --limit 5

The same binary spans four products, so one profile and one command grammar cover Jira, Confluence, Bitbucket, and Jira Service Management. Add --profile prod to target a specific site, and --format json|csv|yaml to shape output for downstream tools. For the full surface, browse the command reference or the Jira CLI hub. Nothing here calls an AI model, so it runs the same in a laptop shell or an unattended pipeline.

FAQ

Related Resources