# Dedalus Labs

> VMs that boot in 250ms. Full Linux, VM-level isolation, persistent storage, scale-to-zero. The production compute platform for AI agents.

## What Dedalus Labs Does

Dedalus Labs builds Dedalus Machines — VMs that boot in 250ms with full Linux, persistent storage, and VM-level isolation. Each machine is a real Linux environment with its own guest kernel, hardware-enforced isolation via Extended Page Tables, and a POSIX-compliant filesystem. Machines scale to zero: sleep is free, wake is instant. The API is three operations: create, exec, sleep.

The platform also provides a unified API gateway for LLM providers, an agent SDK for Python and TypeScript, cloud hosting for MCP servers, and a marketplace for AI tools.

## Key Pages

- [Homepage](https://dedaluslabs.ai): Platform overview and getting started
- [About](https://dedaluslabs.ai/about): Dedalus Machines — architecture, pillars, and how it works
- [Pricing](https://dedaluslabs.ai/pricing): Per-second billing, free tier, Pro and Enterprise plans
- [Marketplace](https://dedaluslabs.ai/marketplace): Browse and discover community MCP servers
- [Blog](https://dedaluslabs.ai/blog): Technical articles and company news
- [Articles](https://dedaluslabs.ai/articles): Practical guides to AI agent infrastructure
- [Documentation](https://docs.dedaluslabs.ai): Developer guides, SDK reference, and API docs
- [Careers](https://dedaluslabs.ai/careers): Open roles at Dedalus Labs

## Products

- **Dedalus Machines**: VMs that boot in 250ms. Full Linux with systemd, VM-level isolation, persistent storage, per-second billing. Sleep is free. Wake is instant. Create, exec, sleep — three API calls.
- **Unified API**: OpenAI-compatible gateway to every model provider. One key, one format, swap models without rewriting code.
- **Agent SDK**: Build production agents in 5 lines. Multi-model routing, streaming, tool use, and automatic retries. Python and TypeScript.
- **Cloud Hosting**: Deploy MCP servers in 3 clicks. No containers, no infrastructure config.
- **Marketplace**: Browse and attach hosted MCP servers to any agent.

## Dedalus Machines — Technical Details

- **Boot time**: 250ms via snapshot restore (not image pull)
- **Isolation**: VM-level. Hardware-enforced via EPT/NPT. Each machine gets its own guest kernel.
- **Storage**: Disaggregated from compute. POSIX-compliant (100% pjdfstest pass rate). Persists across sleep/wake.
- **Billing**: Per-second for active CPU time. $0 when sleeping.
- **Flexibility**: Full Linux with systemd. Any language, any framework, any binary. Root access. GPU support. Nested virtualization.
- **Migrations**: Live VM migrations with zero downtime during fleet maintenance.
- **VMM**: Custom Dedalus Hypervisor built on Cloud Hypervisor.

## Pricing

- **Free tier**: $0/month, 50 free hours of compute, no credit card required
- **Pro ($20/month)**: Higher rate limits, priority support, per-second billing
- **Enterprise (Custom)**: Dedicated support, SLAs, custom limits
- Per-second billing for active compute. $0 when sleeping. Storage priced separately.

## Key Facts

- Founded: 2025
- Headquarters: San Francisco, CA
- Seed round: $11M led by Kindred Ventures and Saga Ventures
- Backed by Y Combinator (S25)
- Core product: Dedalus Machines (VMs)
- Open-source CLI: github.com/dedalus-labs/dedalus-cli

## Frequently Asked Questions

**Q: What are Dedalus Machines?**
A: VMs that boot in 250ms with full Linux, persistent storage, and VM-level isolation. Each machine is a real Linux box with its own guest kernel. The API is three operations: create, exec, sleep.

**Q: How fast do Dedalus Machines boot?**
A: 250ms. State-of-the-art as of 2026. Uses snapshot restore via a custom VMM (Dedalus Hypervisor) built on Cloud Hypervisor. Orders of magnitude faster than existing providers.

**Q: How much do Dedalus Machines cost?**
A: 50 free hours/month. Pro is $20/month. Per-second billing for active CPU time. $0 when sleeping. No imposed session limits.

**Q: What isolation model does Dedalus use?**
A: VM-level isolation — the strongest available. Hardware-enforced via EPT/NPT and privilege ring transitions. Each machine gets its own dedicated guest kernel. Not containers, not user-space kernels.

**Q: How is Dedalus different from E2B or Daytona?**
A: Most providers force a tradeoff: fast startup but restricted environments, or full sandboxes with slow boot. Dedalus Machines give you both: full unrestricted Linux with 250ms boot, VM-level isolation, persistent storage, and scale-to-zero billing.

## Contact

- Website: https://dedaluslabs.ai
- Email: support@dedaluslabs.ai
- GitHub: https://github.com/dedalus-labs
- X: https://x.com/dedaluslabs
- LinkedIn: https://www.linkedin.com/company/dedalus-labs
- Discord: https://discord.gg/RuDhZKnq5R
- YouTube: https://www.youtube.com/@DedalusLabs