# Runcrate

**Source:** https://www.startuphub.ai/organizations/runcrate
**Profile type:** Startup

> Runcrate provides on-demand GPU instances and an OpenAI-compatible inference API for AI teams and agents.

## About

Runcrate is a developer-friendly AI cloud platform that offers on-demand GPU compute and an OpenAI-compatible inference API for over 200 open-source models. It enables AI teams to train, fine-tune, and deploy models with per-second billing, fast cold starts, and persistent storage, all managed from a single account. The platform is designed to remove friction from AI compute, making it accessible and cost-effective for various AI workloads.

## Key facts

- **Slug:** runcrate
- **Website:** https://www.runcrate.ai
- **Founded:** 2024-01-01
- **Employees:** 5
- **Operating status:** Active
- **Sectors / tags:** GPU Cloud, AI Infrastructure, Cloud Computing, Developer Tools, MLOps
- **Social:**
  - LinkedIn: https://www.linkedin.com/company/runcrate
  - Twitter: https://x.com/runcrate
  - GitHub: https://github.com/runcrateai

## Pricing

| Product | Plan | Price | Includes |
| --- | --- | --- | --- |
| Runcrate Platform | Inference | Not listed | Pay per token, Pay per image, Pay per second |
| Runcrate Platform | Compute | Not listed | GPU instances billed for the time they run, Billed by the second on most GPUs, Billed by the minute on a few GPUs |
| Runcrate Platform | Storage | Not listed | Volumes at local-disk speed, Attach to an instance in any region |
| Runcrate Platform | Clusters | Not listed | Reserved multi-node clusters, Prepay 1 to 12 months, Custom quotes available |
| Inference | Text models | $0.0305 / 1M input tokens | - |
| Inference | Vision models | $0.063 / 1M input tokens | - |
| Inference | Image models | $0.00315 / image | - |
| Inference | Video models | $0.0147 / second | - |
| Inference | Speech models | $0.0651 / 1k characters | - |
| Inference | Transcription models | $0.00315 / minute | - |
| Compute GPU instances and crates | H100 GPU, 80GB | $3.55 / hr | - |
| Compute GPU instances and crates | H200 GPU, 141GB | $4.63 / hr | - |

One balance for GPUs and inference. GPUs billed for the time they run, models billed per token.

Pricing varies by specific model. For example, Kimi K3 Fast is $4.73 per 1M input tokens. 5 open models are free with a GitHub account, then half price on shared capacity, list price when busy.

Plans as published by Runcrate, checked October 2026.

## StartupHub.ai scores

- **Ahrefs DR:** 7/100
- **Agent Readiness:** B (79/100)
- **Site Speed:** 63/100
- **Data Trust:** 81/100

## Tech stack

Detected on runcrate.ai, last scanned 2026-10-24.

- **Frontend:** Next.js, Tailwind CSS
- **CDN:** Vercel
- **Email provider:** Google Workspace
- **Affiliate program:** Self-hosted affiliate

## JSON-LD

Structured data aligned with this page's `<script type="application/ld+json">` (included when the profile is server-rendered).

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "BreadcrumbList",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "name": "Home",
          "item": "https://www.startuphub.ai"
        },
        {
          "@type": "ListItem",
          "position": 2,
          "name": "Organizations",
          "item": "https://www.startuphub.ai/search?type=Startup"
        },
        {
          "@type": "ListItem",
          "position": 3,
          "name": "Runcrate",
          "item": "https://www.startuphub.ai/organizations/runcrate"
        }
      ]
    },
    {
      "@type": "ProfilePage",
      "@id": "https://www.startuphub.ai/organizations/runcrate#webpage",
      "url": "https://www.startuphub.ai/organizations/runcrate",
      "name": "Runcrate - AI Startup Profile | StartupHub.ai",
      "description": "This is the StartupHub.ai profile of Runcrate. Runcrate is a startup founded in 2024, operating in GPU Cloud and AI Infrastructure. Runcrate provides on-demand GPU instances and an OpenAI-compatible inference API for AI teams and agents. The company has approximately 5 employees.",
      "dateModified": "2026-10-24T19:02:47.094Z",
      "mainEntity": {
        "@type": "Organization",
        "@id": "https://www.startuphub.ai/organizations/runcrate#organization"
      }
    },
    {
      "@type": "Organization",
      "@id": "https://www.startuphub.ai/organizations/runcrate#organization",
      "name": "Runcrate",
      "url": "https://www.runcrate.ai",
      "logo": "https://cdn.startuphub.ai/storage/v1/object/public/entity-images/logos/182dd655-dd4d-40c8-945b-bf43e4dbc4d2-square.webp",
      "image": "https://cdn.startuphub.ai/storage/v1/object/public/entity-images/logos/182dd655-dd4d-40c8-945b-bf43e4dbc4d2-square.webp",
      "description": "Runcrate is a developer-friendly AI cloud platform that offers on-demand GPU compute and an OpenAI-compatible inference API for over 200 open-source models. It enables AI teams to train, fine-tune, and deploy models with per-second billing, fast cold starts, and persistent storage, all managed from a single account. The platform is designed to remove friction from AI compute, making it accessible and cost-effective for various AI workloads.",
      "sameAs": [
        "https://www.linkedin.com/company/runcrate",
        "https://x.com/runcrate",
        "https://github.com/runcrateai"
      ],
      "foundingDate": "2024-01-01",
      "numberOfEmployees": {
        "@type": "QuantitativeValue",
        "value": 5
      },
      "knowsAbout": [
        "GPU Cloud",
        "AI Infrastructure",
        "Cloud Computing",
        "Developer Tools",
        "MLOps"
      ]
    }
  ]
}
```

## Community reviews

Community reviews below were gathered from public sources into one place. Each quote is copied word for word and links to where it was posted.

- **Reddit r/u/runcrate_ai** | runcrate_ai | 2025-12-28 | positive | highlighted endorsement
  > "GPUs and clusters don't have to break your bank.. AI could that takes you from idea to enterprise, build on Runcrate ."
  > Source: https://www.reddit.com/r/u_runcrate_ai/comments/1py5r41/gpus_and_clusters_dont_have_to_break_your_bank/