Augment Code UPDATE: Remote Agent (Cloud, 10 Parallel Instances)
Source: YouTube — WorldofAI, published 2025-05-09 Tool: Augment Agent
Summary
Walkthrough of Augment Code’s new Remote Agent feature: a cloud-hosted version of the Augment agent that runs in an isolated, mirrored dev environment (not on the developer’s laptop) and supports up to 10 parallel agents working simultaneously while the human is offline. Targeted at the “queue and walk away” workflow — assign issues, agents handle PR drafts, tests, documentation debt; the dev reviews when ready. Waitlist-only at the time of recording (augment.new). Same VS Code / JetBrains / Vim editor support, but the agent runs server-side.
Key facts
- Cloud-based: agent runs in an isolated mirrored dev environment, not local
- 10 parallel agents per session
- GitHub integration: PR creation, CI/CD automation, branch protection
- Editor support: VS Code, JetBrains, Vim
- Status at recording: waitlist-only on
augment.new
Why it matters
Remote Agent is Augment’s async / cloud move, putting it directly in the same category as Google Jules, the Codex async pattern, and Claude Code Ultra Plan. Augment had been the interactive IDE-extension leader (#1 OSS on SWE-bench Verified per their launch); Remote Agent extends that to the queue-and-walk-away surface.
The 10 parallel agents number is the wiki’s first concrete data point on horizontal-scaling for a single dev — multi-agent at the individual level rather than the org level (cf. Paperclip’s company-sim framing).
Position in the async cloud agent cluster
| Jules | Codex (cloud) | Augment Remote Agent | Open-SWE | |
|---|---|---|---|---|
| Vendor | OpenAI | Augment Code | LangChain (OSS) | |
| Parallel agents | 1 | 1 | 10 | 1 |
| Free tier | Yes (5/day) | Plus subscription | Waitlist | Yes (BYO model) |
| Local IDE link | No | CLI / Mac app | Yes (mirrored env) | Self-hostable |