| What it does | Cloud computers for AI agents, billed per second
| Pick your AI model based on evaluation results, not intuition
|
|---|
| Stage | Not given
| Not given
|
|---|
| Founded | Not given
| Not given
|
|---|
| Based in | Not given
| US
|
|---|
| Pricing model | Freemium
| Not given
|
|---|
| Pricing | Starter $20/month ($20 usage credit, 5 machines awake), Growth $100/month ($100 usage credit, 25 machines awake), Scale $500/month ($500 usage credit, 60 machines awake), Dedicated $5,000/month ($5,000 usage credit, 500 machines awake). Agent boxes $1/month per 2 GB. Sandboxes from $0.011/hour. Linu
| Not given
|
|---|
| Key features | - Agent boxes with persistent disk that sleep when idle and wake in about a second
- Sandboxes for running untrusted or generated code, deleted when task ends
- Swarms to run thousands of sandboxes on one command for evals and training
- Linux desktops that agents can see, click and type on with human takeover
- Windows desktops for desktop apps with Windows Server 2022 and included licence
- Android phones with AOSP Android 15 for mobile app automation and app installs
- Deploy agents from templates, GitHub repos or Docker images
- Marketplace with pre-configured desktop images and software
- Plan fees convert to usage credit with per-second billing for machines
| - Automatic model routing based on evaluation results
- GitHub repository integration for eval workflows
- Stores evidence of evaluation results
- Automatic re-evaluation when new models are released
- Dashboard for monitoring routes
- MCP server integration with ChatGPT, Claude, and Cursor
- Configurable evaluation checks and criteria
|
|---|
| What makes it different | - Per-second billing instead of hourly, with plan fees that convert to usage credit
- Agent boxes cost $1/month per 2 GB awake or asleep, enabling persistent agents at low cost
- Support for computer use with visual interfaces (Linux, Windows, Android)
- Swarms enable efficient batch processing and evaluation at scale
| - Test your own LLM applications against real models automatically
- Routes requests to the best-performing model in real time
- Automatically updates routing when new models are released
- Keeps your eval results and configurations under your control
|
|---|
| Community votes | 0 | 0 |
|---|
| DR (Domain Rating by Ahrefs) | 28 ▲ +1 (up) since last check | Not checked yet ● no change since last check |
|---|
| Trust Flow (Majestic) | 10 | 0 |
|---|
| Citation Flow (Majestic) | 33 | 0 |
|---|
| Referring domains (Majestic) | 214 | 0 |
|---|
Maritime
BenchRouter