---
title: "Run isolated AI agents at scale through one API — AgentSky"
description: "Start one isolated cloud agent per task through a stable API. You keep queue, budget, and output policy while AgentSky operates sandboxes, events, and recovery."
url: https://agentsky.dev/use-cases/run-agents-at-scale
---

[All use cases](https://agentsky.dev/use-cases)

# Run an agent fleet without building one

Start one isolated cloud agent per task with a single API call.

[Read the developer API](https://agentsky.dev/developers) [See the factory architecture](https://agentsky.dev/software-factories)

## Starter prompt

```
Here are [N] items to work through: [list, file, or query].
Run one agent per item, apply the same instructions to each, and give me a single table of results with the failures separated out.
```

Running many agent tasks in parallel means operating infrastructure teams never signed up to build.

ScaleHermes + DeepSeek V4.1 Flash

## Related

- [Hermes in the catalog](https://agentsky.dev/agents/hermes)
- [Read the reference architecture](https://agentsky.dev/software-factories)
- [Run it from the API](https://agentsky.dev/developers)
- [What a run costs](https://agentsky.dev/pricing)

## Choose something else when

Own the execution layer when you need self-hosting, custom queue placement, GPU training, kernel-level controls, or workload guarantees outside AgentSky's contract.

## Best for

- Product teams starting many isolated agent tasks through one API while retaining control of task routing, budgets, and downstream outputs.
- You bring the task list and instructions.
- You bring the agent stack per task.
- You bring the concurrency and budget.

## How it works

1. Step 1 One API call, one isolated agent per task Create sessions programmatically — each task gets its own cloud sandbox, isolated from every other run. Priced per session, not per seat or per machine.
2. Step 2 Right agent stack routes to each task Route different task types to different harness and model combinations. The API contract stays the same regardless of what runs behind it — the fleet is heterogeneous by default.
3. Step 3 Structured events stream back per session Stream events per session. Artifacts, logs, and history land per run — consistent format across every agent in the fleet, ready to aggregate or route downstream.

## What AgentSky runs

- A cloud sandbox per session
- Streaming events
- Lifecycle and recovery across the fleet

## Software factories

AgentSky's software factory pattern creates one isolated cloud agent per task through the Agent API. The product owns queue logic and output routing; AgentSky operates every sandbox, event stream, and recovery layer.

A reference architecture, not a capacity or throughput guarantee; concurrency, budget, and workload fit still require validation.

[Read the reference architecture](https://agentsky.dev/software-factories)

---

Markdown copy of https://agentsky.dev/use-cases/run-agents-at-scale. Every agentsky.dev page has one: add `.md` to its address or send `Accept: text/markdown`. Start from https://agentsky.dev/llms.txt.
