TensorDock

aid: tensordock
url: https://raw.githubusercontent.com/api-evangelist/tensordock/refs/heads/main/apis.yml
apis:
  - aid: tensordock:tensordock-marketplace-api
    name: TensorDock Marketplace API
    tags:
      - GPU
      - Cloud
      - Compute
      - Marketplace
      - Virtual Machines
    humanURL: https://documenter.getpostman.com/view/20973002/2s8YzMYRDc
    properties:
      - url: https://documenter.getpostman.com/view/20973002/2s8YzMYRDc
        type: Documentation
      - url: openapi/tensordock-marketplace-api-openapi.yml
        type: OpenAPI
    description: >-
      The original TensorDock Marketplace REST API (v0) for deploying, managing, starting, stopping, and deleting GPU
      virtual machines across the global marketplace of independent hardware hosts. Endpoints accept api_key + api_token
      authorization pairs as form parameters and cover hostnode discovery, on-demand and spot instance deployment,
      instance lifecycle, container scaling, and billing balance lookups.
  - aid: tensordock:tensordock-instances-api
    name: TensorDock Instances API
    tags:
      - GPU
      - Cloud
      - Compute
      - Instances
      - Virtual Machines
    humanURL: https://dashboard.tensordock.com/api/docs/instance-management
    properties:
      - url: https://dashboard.tensordock.com/api/docs/instance-management
        type: Documentation
      - url: https://dashboard.tensordock.com/api/docs/instance-creation
        type: Documentation
      - url: openapi/tensordock-instances-api-openapi.yml
        type: OpenAPI
      - url: json-schema/tensordock-instance-schema.json
        type: JSONSchema
      - url: json-ld/tensordock-context.jsonld
        type: JSONLD
    description: >-
      The current TensorDock Instances API (v2) at https://dashboard.tensordock.com/api/v2/instances for creating,
      listing, inspecting, starting, stopping, modifying, and deleting GPU and CPU virtual machines. Uses Bearer token
      authentication and a JSON:API-style envelope of `data.type`, `data.id`, and `data.attributes`. Supports both
      location-based and hostnode-based deployment patterns.
  - aid: tensordock:tensordock-secrets-api
    name: TensorDock Secrets API
    tags:
      - GPU
      - Cloud
      - Secrets
      - SSH Keys
      - Credentials
    humanURL: https://dashboard.tensordock.com/api/docs/secrets
    properties:
      - url: https://dashboard.tensordock.com/api/docs/secrets
        type: Documentation
      - url: openapi/tensordock-secrets-api-openapi.yml
        type: OpenAPI
    description: >-
      The TensorDock Secrets API (v2) for managing SSH keys and generic secrets that are encrypted at rest and in
      transit and can be attached to instances at deploy time. Two secret types are supported — `SSHKEY` for SSH
      authentication and `SECRET`/`GENERIC` for application credentials whose values are not returned after creation.
name: TensorDock
tags:
  - GPU
  - Cloud
  - Marketplace
  - Compute
  - Virtual Machines
  - AI
  - Machine Learning
  - Bare Metal
  - Spot Instances
  - Containers
kind: contract
image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/apis-json-logo.jpg
access: 3rd-Party
common:
  - type: Portal
    url: https://www.tensordock.com
  - type: Documentation
    url: https://docs.tensordock.com
  - type: Documentation
    url: https://dashboard.tensordock.com/api/docs
    name: API Documentation
  - url: https://documenter.getpostman.com/view/20973002/2s8YzMYRDc
    name: Postman API Documentation
    type: Documentation
  - url: https://dashboard.tensordock.com/api/docs/getting-started
    name: Getting Started
    type: GettingStarted
  - url: https://dashboard.tensordock.com
    name: TensorDock Dashboard
    type: Console
  - url: https://dashboard.tensordock.com/deploy
    name: Deploy GPU VM
    type: Sandbox
  - url: https://dashboard.tensordock.com/deploy_cpu
    name: Deploy CPU VM
    type: Sandbox
  - url: https://dashboard.tensordock.com/api
    name: API Keys
    type: SignUp
  - url: https://blog.tensordock.com
    name: TensorDock Blog
    type: Blog
  - url: https://tensordock.com/security
    name: Security and Reliability
    type: TrustCenter
  - url: https://docs.tensordock.com/legal-information/legal-information
    type: Documentation
    name: Legal Information
  - url: https://docs.tensordock.com/legal-information/terms-of-service-tos
    type: TermsOfService
  - url: https://docs.tensordock.com/legal-information/tensordock-privacy-policy
    type: PrivacyPolicy
  - url: https://docs.tensordock.com/legal-information/acceptable-use-policy-aup
    type: TermsOfService
    name: Acceptable Use Policy
  - url: https://docs.tensordock.com/legal-information/downtime-compensation
    type: Documentation
    name: Downtime Compensation
  - url: https://docs.tensordock.com/legal-information/taxes-vat-gst
    type: Documentation
    name: Taxes, VAT, GST
  - url: https://marketplace.tensordock.com/support
    name: Support
    type: Support
  - url: https://marketplace.tensordock.com/faq
    name: FAQ
    type: Support
  - url: https://tensordock.userjot.com/
    name: User Suggestions
    type: Forum
  - type: Twitter
    url: https://twitter.com/TensorDock
  - type: LinkedIn
    url: https://www.linkedin.com/company/tensordock
  - type: Instagram
    url: https://instagram.com/tensordock
  - type: YouTube
    url: https://www.youtube.com/channel/UCsCAK6krPmRe3Y7Hp5QHXQg
  - type: Forum
    url: https://discord.gg/Xyzjjuj6zf
    name: TensorDock Discord
  - type: GitHubOrganization
    url: https://github.com/tensordock
  - url: https://github.com/tensordock/tensordock-rs
    name: Rust API Wrapper
    type: SDK
  - url: https://pypi.org/project/tensordock/
    name: Unofficial Python SDK
    type: SDK
  - url: https://github.com/tensordock/dashboard-frontend-template
    name: Whitelabel Storefront Template
    type: Tool
  - url: https://github.com/tensordock/wireguard-manager
    name: WireGuard Manager
    type: Tool
  - url: https://github.com/tensordock/tensorblog
    name: TensorBlog
    type: Tool
  - url: https://github.com/tensordock/hosting-install
    name: Hosting Install Scripts
    type: Tool
  - url: https://docs.tensordock.com/virtual-machines/introduction-to-core-compute-vms
    name: Introduction to Core Compute VMs
    type: Documentation
  - url: https://docs.tensordock.com/virtual-machines/spot-instances
    name: Spot Instances
    type: Documentation
  - url: https://docs.tensordock.com/virtual-machines/cloud-init
    name: cloud-init
    type: Documentation
  - url: https://docs.tensordock.com/virtual-machines/how-to-ssh-into-your-instance
    name: SSH Access Guide
    type: Documentation
  - url: https://docs.tensordock.com/virtual-machines/how-to-rdp-into-your-instance
    name: RDP Access Guide
    type: Documentation
  - url: https://docs.tensordock.com/whitelabel/overview
    name: Whitelabel Overview
    type: Documentation
  - url: https://docs.tensordock.com/whitelabel/setting-up-a-storefront
    name: Setting Up a Storefront
    type: Documentation
  - url: https://www.tensordock.com/host.html
    name: Become a Host
    type: SignUp
  - url: plans/tensordock-plans-pricing.yml
    type: Plans
  - url: rate-limits/tensordock-rate-limits.yml
    type: RateLimits
  - url: finops/tensordock-finops.yml
    type: FinOps
  - type: Pricing
    url: https://marketplace.tensordock.com
    data:
      - id: gpu-h100-sxm5
        name: NVIDIA H100 SXM5 80GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 2.25
            metric: gpu
            timeFrame: hour
            description: Per-GPU hourly rate from $2.25.
        description: Flagship Hopper architecture GPU for large-scale AI training and inference.
      - id: gpu-a100-sxm4-80gb
        name: NVIDIA A100 SXM4 80GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.87
            metric: gpu
            timeFrame: hour
        description: 80GB Ampere SXM4 for AI training and HPC workloads.
      - id: gpu-a100-pcie-80gb
        name: NVIDIA A100 PCIe 80GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 1.27
            metric: gpu
            timeFrame: hour
        description: 80GB Ampere PCIe variant.
      - id: gpu-rtx-5090
        name: NVIDIA RTX 5090 32GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.56
            metric: gpu
            timeFrame: hour
        description: Blackwell-architecture consumer GPU with 32GB VRAM.
      - id: gpu-rtx-4090
        name: NVIDIA RTX 4090 24GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.32
            metric: gpu
            timeFrame: hour
        description: Ada Lovelace 24GB consumer GPU.
      - id: gpu-rtx-3090
        name: NVIDIA RTX 3090 24GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.17
            metric: gpu
            timeFrame: hour
        description: 24GB Ampere consumer GPU.
      - id: gpu-rtx-6000-ada
        name: NVIDIA RTX 6000 Ada 48GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.72
            metric: gpu
            timeFrame: hour
        description: 48GB Ada Lovelace workstation GPU for rendering and ML.
      - id: gpu-l40s
        name: NVIDIA L40S 48GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.49
            metric: gpu
            timeFrame: hour
        description: 48GB Ada datacenter GPU.
      - id: gpu-l4
        name: NVIDIA L4 24GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.24
            metric: gpu
            timeFrame: hour
        description: 24GB Ada efficiency GPU.
      - id: gpu-a6000
        name: NVIDIA RTX A6000 48GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.42
            metric: gpu
            timeFrame: hour
        description: 48GB Ampere workstation GPU.
      - id: gpu-v100-sxm3
        name: NVIDIA V100 SXM3 32GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.3
            metric: gpu
            timeFrame: hour
        description: 32GB Volta SXM3 GPU.
      - id: gpu-v100-sxm2
        name: NVIDIA V100 SXM2 16GB
        type: GPU
        entries:
          - geo: Global
            unit: 1
            label: Hourly
            price: 0.21
            metric: gpu
            timeFrame: hour
        description: 16GB Volta SXM2 GPU.
      - id: cpu-compute
        name: CPU Compute
        type: CPU
        entries:
          - geo: Global
            unit: 1
            label: Per vCPU
            price: 0.012
            metric: vcpu
            timeFrame: hour
        description: Intel Xeon and AMD EPYC virtual CPUs from $0.012/vCPU/hour. Ideal for transcoding and batch processing.
      - id: spot
        name: Spot Instances
        type: Pricing Model
        entries:
          - geo: Global
            unit: 1
            label: Bid Price
            price: Variable
            metric: gpu
            timeFrame: hour
        description: Interruptible instances with custom bid pricing for opportunistic workloads.
      - id: reserved
        name: Reserved Pricing
        type: Pricing Model
        entries:
          - geo: Global
            unit: 1
            label: Contact
            price: Call
            metric: gpu
            timeFrame: month
        description: Volume reservations available upon request — contact sales for committed-use discounts.
    name: Pricing
  - type: Features
    data:
      - GPU marketplace aggregating 30,000+ GPUs across 100+ locations in 20+ countries
      - >-
        42+ NVIDIA GPU models from consumer (RTX 3090/4090/5090) to data center (H100 SXM5, A100, V100, L40S, L4, RTX
        6000 Ada, A6000)
      - Per-microsecond billing with pay-as-you-go and reserved pricing on request
      - On-demand virtual machines with KVM virtualization, root access, and dedicated GPU passthrough
      - Spot/interruptible instances with custom bid pricing and soft-validate APIs
      - Instant VMs (beta) with sub-30-second deployment from pre-configured templates
      - Container deployment, scaling, and termination endpoints
      - CPU-only Compute VMs from $0.012/vCPU/hour for transcoding and batch jobs
      - Windows 10 and Linux (Ubuntu, Debian, CentOS) images plus custom cloud-init support
      - Whitelabel storefront product for resellers and managed cloud providers
      - >-
        Two REST APIs — Marketplace v0 (api_key + api_token, form parameters) and Instances v2 (Bearer token, JSON:API
        envelope)
      - Secrets API (v2) with SSHKEY and GENERIC/SECRET types, encrypted at rest and in transit
      - Hostnode discovery API listing all available GPU/CPU configurations per location
      - 99.99% uptime SLA per host with agent-based monitoring and downtime compensation policy
      - SSH access revoked from hosts and active monitoring for tenant isolation
      - >-
        Postman collection covering Authorization, Hostnodes, Virtual Machines, Spot validation, Containers, and Billing
        endpoint groups
      - Rust client wrapper (tensordock-rs) and community Python SDK
      - 100 requests per minute per organization on the v2 API
      - Customer balance, revenue, and monthly summary endpoints for host operators
    sources:
      - https://www.tensordock.com
      - https://docs.tensordock.com
      - https://dashboard.tensordock.com/api/docs
      - https://documenter.getpostman.com/view/20973002/2s8YzMYRDc
      - https://blog.tensordock.com
    updated: '2026-05-25'
created: '2026-05-25'
modified: '2026-05-25'
position: Consuming
description: >-
  TensorDock operates a global GPU cloud marketplace that connects independent hardware hosts with customers needing
  affordable on-demand and spot GPU compute. The platform exposes two REST APIs — the original Marketplace API (v0) at
  marketplace.tensordock.com and the newer Instances API (v2) at dashboard.tensordock.com — to deploy, manage, and scale
  virtual machines across 100+ locations in 20+ countries with 42+ GPU models ranging from consumer RTX cards to H100
  SXM5. TensorDock targets AI startups, researchers, rendering shops, and gaming services that need cost-effective
  compute (advertised at up to 80% less than hyperscalers) without quotas or long-term commitments.
maintainers:
  - FN: Kin Lane
    email: info@apievangelist.com
    X: apievangelist
    url: https://apievangelist.com
specificationVersion: '0.16'