Cloud agents

Gain visibility into the pull requests created by AI agents running in the cloud. See how they perform across metrics like throughput, merge percentage, and batch size.

What is a cloud agent?

Cloud agents are AI coding agents responsible for creating pull requests end-to-end — planning work, editing code, running tests, and responding to reviews — not just suggesting code in your editor.

Most often, users assign AI cloud agents individually from issue trackers or prompt them to create a single pull request. AI cloud agents can also be autonomous, in which case they will have a backlog or objective and continuously pick work items on their own.

Built-in cloud agents

GitHub Copilot cloud agent can create pull requests from various sources, including IDEs, GitHub.com, GitHub issues, GitHub Mobile, CLI, and MCP.
Cursor Cloud Agents can create pull requests from cursor.com/agents, or you can trigger them via integrations, such as Slack and Linear.
Claude Code GitHub Actions can create pull requests when you mention @claude in PRs or issues.
Claude Code on the web is available in research preview to selected pricing plans and can create pull requests from claude.ai/code.
Claude by Anthropic GitHub partner agent can create PRs from github.com, GitHub Mobile, and VS Code.
Codex cloud can create pull requests from chatgpt.com/codex, the Codex CLI and IDE extension, and integrations such as GitHub, Slack, and Linear.
CodeRabbit is primarily a review agent, but it can also open a pull request directly, for example to apply a fix it suggested during review. Swarmia attributes those PRs as CodeRabbit cloud agent PRs.
Other review agents, such as Cubic, Gemini Code Assist, Graphite, Greptile, Macroscope, and Qodo, are attributed the same way when they open a pull request. See the full list of built-in review agents.
Devin can create pull requests from its web app and integrations such as Slack and Linear. See Connect Devin to Swarmia for optional usage data.

Don't see your tool? Add it as a custom AI tool to get the same detection.

Cloud agent PRs in Swarmia

In Swarmia, pull requests authored by AI tools are attributed to the people who initiated them. For example, if GitHub Copilot creates a PR on your behalf, Swarmia shows you as its author.

The AI coding agent used to create a pull request is shown on top of the pull request author’s avatar.

Swapping the author affects the whole app: the pull request counts toward the person's team, their contribution stats, and their work log. The agent stays visible through the badge on the author's avatar and the cloud agent AI tool tag. Commits keep their original author, which is what the agent commit percentage chart measures.

Understanding your cloud agent performance

To help you understand how your AI cloud agents are performing, Swarmia automatically detects pull requests created by any of the cloud agents above, as well as any custom AI tool you've linked a contributor to. Swarmia provides a dedicated cloud agents (previously called "coding agents") view for analyzing them.

Agent PR throughput

The Agent PR throughput chart shows how much work AI cloud agents help you ship.

Agent PRs, Merged, and Closed show the raw number of AI cloud agent pull requests.

Merge percentage tells you about your conding agent performance and how effectively people are able to use agents. Ideally, the number should grow close to 100%, but never quite reach it. Even though every closed pull request is technically waste (i.e., work done that was never shipped), a small percentage of closed pull requests indicates that people are experimenting with AI cloud agents and pushing their limits.

Agent commit percentage

The Agent commit percentage chart breaks down your merged pull requests based on how much human intervention was needed for the work to be shipped.

Ideally, all your agent pull requests would have 100% agent commits, which means that the person was able to complete the whole task through the agent interface (e.g., GitHub Copilot's pull request integration). If a pull request ends up in the other buckets, it means that someone needed to pull up their coding environment and commit code to finish the work.

Batch size

The Batch size chart helps you compare the distribution of AI cloud agent pull requests with your overall Batch size metrics. It also gives you an indication of the complexity of work AI cloud agents can tackle.

Depending on how you start using AI cloud agents, you might see that most agent pull requests initially fall in the smallest 1–5 or largest 500+ buckets. As your AI cloud agent usage develops, you should see the distribution move toward the buckets in the middle. This means your agents can tackle tasks of all sizes and ship code in manageable increments.

Teams breakdown

The "Teams" tab provides a breakdown of the core metrics shown in the top charts, grouped by team.

Pull requests tab

You can use the "Pull requests" tab in combination with the pull request filters at the top to inspect the individual agent pull requests.

Filtering PRs from cloud agents

You can use cloud agent mode in the AI tool filter to view pull requests created by cloud agents across the Swarmia app.

Frequently asked questions

Which AI cloud agents does Swarmia support?

We have built-in support for the cloud agents listed above. For any other cloud agent, or one you've built yourself, you can add it as a custom AI tool and link its contributor. Swarmia then attributes its pull requests to it with high confidence, the same way it recognizes the built-in agents.

How do you identify PRs created with cloud agents?

Swarmia attributes a PR to a cloud agent if the PR or its first commit has been authored by one of the cloud agents listed above, or a contributor linked to a custom AI tool.

Read more about our automatic AI tool detection.

Who is the author of a cloud agent PR?

The person who started the agent, not the agent's bot account. Swarmia identifies them from the commit co-authors, or from the person the agent assigned to the pull request. Read more about how Swarmia finds the person behind an agent PR.

Why does a cloud agent PR show the agent as the author instead of a person?

Swarmia couldn't find the human behind it. That happens when:

The agent neither credited a co-author on the first two commits nor assigned anyone to the pull request.
The only human co-author appears on the third commit or later.
The pull request is a GitLab merge request, where Swarmia relies on commit co-authors alone.

Why is my PR not attributed to a cloud agent?

Swarmia attributes a PR to a cloud agent if the PR or its first commit has been authored by one of the cloud agents listed above.

Example 1: With GitHub Copilot coding agent, GitHub shows Copilot as the PR author. Swarmia shows this as a GitHub Copilot agent PR:

Note: Regardless of what GitHub displays, Swarmia always shows the author as the person who initiated the AI coding agent PR.

Example 2: When using Cursor Cloud Agents, GitHub shows cursoragent as the author of the first commit, and the user who started the agent as the PR author. Swarmia shows this as a Cursor agent PR:

Example 3: The first commit in this PR was authored by Claude but committed by cursoragent. Swarmia shows this as a Claude Code agent PR:

If I ask an AI tool to create a PR with my local changes, is it considered a cloud agent PR?

No, we don't regard that as a cloud agent PR since you review and accept the suggestions (and perhaps manually write some of the code) before creating the PR.

The AI tool uses your local Git identity, so you are marked as the PR author and the commit author. (The tool might mark itself as the co-author, however.)

Swarmia shows these PRs as AI-assisted.

How does Swarmia attribute GitHub partner agents?

The Claude by Anthropic and Codex by OpenAI GitHub partner agents can create PRs from github.com, GitHub Mobile, and VS Code. Although you pay for them through the GitHub Copilot billing, Swarmia attributes them as the Claude Code and Codex cloud agents, because that's the harness each one runs.

How does Swarmia find the people behind custom agent PRs?

If a PR is opened by a (bot) contributor linked to a custom AI tool, Swarmia looks for the person who started the agent in two places:

Commit co-authors. The first non-bot author credited in a Co-authored-by trailer on the pull request's first two commits. Swarmia checks two commits because some agents start with a planning commit that has no co-author.
Assignment by the agent. If no co-author turns up, Swarmia looks for the first person the agent itself assigned to the pull request. Assignments made by anyone other than the agent that opened the PR are ignored, and this signal is available on GitHub only.

If neither finds a human, the pull request stays attributed to the agent's bot account.

What if there isn't a single person behind my cloud agent PR?

Autonomous agents might create pull requests without a person to attribute them to. These pull requests stay attributed to the agent's bot account and don't appear in team-scoped views unless the bot belongs to a team.

To analyze this work, you can create an Agents team and add the bot contributor to it. The pull requests will then appear in that team's metrics and other team-scoped views. Bots can be team members, just like people.