Skip to content
Agents tracked: 157 Downloads (7d): 216M up 9.6% GitHub stars: 4.2M VS Code installs: 145M Releases (7d): 233 Agent pull requests (last week): 940K Updated Oct 6, 2026

Cloud coding agents

Agents that take a task, work on it in a remote sandbox and come back with a pull request for you to review.

# Agent Pulse Downloads 7d 7d 30d VS Code installs Stars Latest release Price Last 90 days
31
DevinCognition
53 — — — — — — $20/mo
121
JulesGoogle (Google Labs)
19 2,765 up 13.2% up 6.7% — — — Free + $19.99/mo

Ranks are overall positions by Pulse Score. Click a column to sort by what matters to you. Methodology.

How they differ from local agents

A cloud agent clones your repository into its own environment, so it can run for a long time, in parallel with others, without tying up your machine. You assign work from a chat, an issue tracker or a code host and review the result as a pull request.

The trade-off is context. The agent only knows what is in the repository and whatever setup script you give it, so tasks that depend on local services, secrets or tacit knowledge go less well.

A signal that only exists for this group

Because these agents open pull requests under their own bot accounts, their public output can be counted. The GitHub activity page tracks weekly pull request volume and the share that gets merged.

Head-to-head comparisons

Related guides

Questions about cloud coding agents

What does merge rate tell me about a cloud coding agent?

It is the share of the agent's public pull requests that were accepted. A higher rate suggests more of its work is usable as delivered, but the rate also depends on how the agent is used: one that only opens a pull request when asked will score higher than one that opens them speculatively.