# Islo

> Islo deploys and runs autonomous coding agents in secure environments connected to your tools, so they can finish real engineering work.

Islo gives each agent a persistent computer with your stack, scoped tool access, gateway credential injection, and verification against your real services. Agents keep working after you disconnect and return finished work with evidence you can inspect.

- **Works unattended** - Agents keep working after you close your laptop
- **Connected to your tools** - Scoped access to GitHub, Slack, Linear, Jira, databases, and internal services
- **Real development environments** - The same repo, dependencies, services, browser, configuration, and environment variables as a developer machine
- **Gateway controls** - Credentials injected at egress, scoped network policies, response filtering, and per-request audit
- **Verification** - End-to-end checks confirm that agents completed the workflow
- **Isolation** - Each agent runs on its own computer

Quick start:

```bash
curl https://islo.dev/install.sh | bash
islo login
islo use --agent claude
islo use --agent claude --task "Refactor payments to use Stripe webhooks"
islo use swift-lantern
```

SDK quick start:

```bash
npm i @islo-labs/sdk
```

```typescript
import { Islo } from "@islo-labs/sdk"

const islo = new Islo()
const sandbox = await islo.sandboxes.createSandbox({ name: "agent-work" })
await islo.sandboxes.execInSandbox({
  sandboxName: sandbox.name,
  command: ["agent", "run", "--task", "Fix flaky checkout test"],
})
```

Pay only for what you use: CPU $0.07/CPU-hour, memory $0.04/GB-hour, storage $0.0007/GB-hour, with credits to start and no card required.

## Questions

- [What is Islo?](https://islo.dev/): Deploys and runs autonomous coding agents in secure environments connected to your tools.
- [Can coding agents really work autonomously?](https://islo.dev/): Each agent runs on a persistent computer with scoped tool access and keeps working after you disconnect.
- [How can I trust autonomous agents on Islo?](https://islo.dev/): Start with one workflow, verify against your stack, and expand with evals, gateway policies, and audit.
- [How can coding agents test a product like a real user?](https://islo.dev/simulation-agents): Simulation agents boot the full stack, use a browser, and return pass/fail evidence.
- [Can an agent review a pull request against a running stack?](https://islo.dev/pr-review): Loads related pull requests on one computer, runs tests and services, and posts findings to GitHub.
- [Can an agent turn a Linear or Jira bug into a pull request?](https://islo.dev/bug-to-pr): Gives an agent the ticket, repo, services, and browser it needs to reproduce, fix, test, and open a PR.
- [Can an agent investigate an alert and open a fix?](https://islo.dev/alert-monitoring): Inspects logs and metrics, traces the failure, and opens a fix pull request when the cause is in the codebase.
- [How should I run LLM-generated code safely?](https://islo.dev/untrusted-code): Runs generated scripts on disposable isolated computers with scoped egress.
- [Islo or Cursor Cloud Agents?](https://islo.dev/compare/cursor-cloud-agents): Cursor fits quick browser sessions; Islo fits long-running work that needs your environment and tools.
- [Islo or Claude Code for Web?](https://islo.dev/compare/claude-code-web): Claude Code for Web fits quick browser help; Islo fits unattended work with your environment and scoped credentials.
- [What is the best sandbox for Claude Code?](https://islo.dev/blog/competition-docker-daytona-e2b-sprites-coder): Depends on local isolation, cloud execution, AWS platforms, or complete unattended workflows.
- [When should I use Kubernetes Agent Sandbox?](https://islo.dev/blog/agent-sandbox-why-should-i-use-it): When your platform team wants agent environments as Kubernetes resources.
- [What use cases does Islo support?](https://islo.dev/use-cases): Simulation agents, PR review, bug to PR, alerts, untrusted code, evals, and migrations.
- [What is a software factory?](https://islo.dev/software-factory): A repeatable system that moves engineering work from intake through implementation, independent review, end-to-end verification, and human approval.
- [How do Islo sandboxes differ from containers?](https://islo.dev/sandboxes): Isolated computers that stay up for hours, keep state between runs, and let you reattach mid-task.
- [How do agents use our tools without holding credentials?](https://islo.dev/gateway): The gateway injects tokens at egress and logs every outbound request.
- [What is an agent sandbox?](https://islo.dev/learn/agent-sandbox): An isolated computer where agents run code, browse, and hit real services under policy.
- [What are agent evals?](https://islo.dev/agent-evals): Define tasks, run graded attempts in sealed computers, and compare models before widening access.
- [Harbor and RL infrastructure on Islo?](https://islo.dev/rl): Parallel benchmark rollouts with gateway-enforced anti-cheating controls.
- [Islo for startups?](https://islo.dev/solutions/startups): Run a software factory before you hire a platform team.
- [Islo for enterprise?](https://islo.dev/solutions/enterprise): Deploy autonomous agents with evals, BYOC, and exportable audit.
- [How does Islo compare to Devin?](https://islo.dev/compare/devin): Infrastructure vs Devin's bundled autonomous engineer product.
- [How does Islo compare to Factory?](https://islo.dev/compare/factory): Infrastructure vs Factory's integrated software factory product.
- [How does Islo compare to E2B?](https://islo.dev/compare/e2b): Gateway policies, credential injection, and orchestration beyond isolated sandboxes.
- [How does Islo compare to Cursor Cloud Agents?](https://islo.dev/compare/cursor-cloud-agents): BYOC, harness choice, and gateway controls.
- [Should we build agent infra ourselves?](https://islo.dev/compare/diy): Compare Islo to rolling your own stack with containers and a homegrown gateway.
- [Does Islo integrate with GitHub and Linear?](https://islo.dev/integrations): Built-in integrations for GitHub, Linear, Notion, Atlassian, Slack, Anthropic, Cursor, OpenAI, AWS, and GCP.
- [What is Islo pricing?](https://islo.dev/pricing): Pay-as-you-go CPU, memory, and storage pricing with a cost calculator.
- [Is Islo secure for enterprise use?](https://islo.dev/security): Isolation, gateway credential injection, audit, and BYOC.
- [Agent evals and benchmarks on Islo?](https://islo.dev/agent-evals): Run eval suites with gateway anti-cheat controls.
- [Does Islo run a community?](https://islo.dev/community): Autonomous Engineering talks for engineers running agents in production.

## Links

- [Homepage](https://islo.dev/): Product overview and quick start
- [Software factory](https://islo.dev/software-factory): Ticket-to-PR automation with verification and human approval
- [Sandboxes](https://islo.dev/sandboxes): Persistent agent computers with your real stack
- [Gateway](https://islo.dev/gateway): Credential injection, egress policy, and audit
- [Use cases](https://islo.dev/use-cases): Simulation, PR review, bug to PR, alerts, evals, and migrations
- [Solutions hub](https://islo.dev/solutions): Startups, enterprise, and Databricks
- [Solutions for startups](https://islo.dev/solutions/startups): Ship a software factory without a platform team
- [Solutions for enterprise](https://islo.dev/solutions/enterprise): Deploy with evals, BYOC, and audit
- [Customers](https://islo.dev/customers): Customer stories
- [Community talks](https://islo.dev/community): Recorded Autonomous Engineering sessions
- [Compare](https://islo.dev/compare): Competitor comparisons
- [Integrations](https://islo.dev/integrations): GitHub, Linear, Slack, and more
- [Pricing](https://islo.dev/pricing): Pay-as-you-go pricing and calculator
- [Security](https://islo.dev/security): Isolation, gateway, and BYOC
- [Learn glossary](https://islo.dev/learn): Agent infrastructure terms
- [Blog](https://islo.dev/blog): Engineering posts and comparisons
- [Agent evals](https://islo.dev/agent-evals): Graded agent evaluation workflow
- [Harbor / RL](https://islo.dev/rl): Parallel benchmark infrastructure with anti-cheating controls
- [Docs](https://docs.islo.dev/overview): Product documentation
- [Dashboard](https://app.islo.dev): Start an agent
- [Agent context](https://islo.dev/AGENTS.md): Machine-readable product context
- [RSS feed](https://islo.dev/rss.xml): Blog feed
- [Sitemap](https://islo.dev/sitemap.xml): Site index
- [GitHub](https://github.com/islo-labs): Open source and SDK
- [X / Twitter](https://x.com/islolabs): Product updates

## Optional

- [Contact](mailto:support@islo.dev): Questions or support