# Thunder Compute - [Quickstart](https://www.thundercompute.com/docs/mcp-server/quickstart.md): Connect AI tools to GPUs to develop faster - [Technical Specifications](https://www.thundercompute.com/docs/technical-specs.md): Hardware specifications, networking details, and pre-installed software for Thunder instances - [Compatibility](https://www.thundercompute.com/docs/compatibility.md): Review known Thunder compatibility limitations, including managed memory, CUDA MPS, FFmpeg, and GPU profilers. - [Authentication](https://www.thundercompute.com/docs/mcp-server/operations/authentication.md): Use your AI agent to authenticate with Thunder, or switch to another interface for direct control. - [Creating Instances](https://www.thundercompute.com/docs/mcp-server/operations/creating-instances.md): Ask your AI agent to create Thunder instances, or switch to another interface for direct control. - [Connecting to Instances](https://www.thundercompute.com/docs/mcp-server/operations/connecting-to-instances.md): Ask your AI agent to connect to instances and run commands, or switch to another interface for direct control. - [Monitoring Instances](https://www.thundercompute.com/docs/mcp-server/operations/monitoring-instances.md): Ask your AI agent for instance status and usage, or switch to another interface for direct control. Monitor performance, debug issues, and manage resources. - [File Transfers](https://www.thundercompute.com/docs/mcp-server/operations/file-transfers.md): Ask your AI agent to move data to and from Thunder instances, or switch to another interface for direct control. - [Port Forwarding](https://www.thundercompute.com/docs/mcp-server/operations/port-forwarding.md): Ask your AI agent to set up or inspect port forwarding, or switch to another interface for direct control. - [Modifying Instances](https://www.thundercompute.com/docs/mcp-server/operations/modifying-instances.md): Ask your AI agent to update instance configuration, or switch to another interface for direct control. - [Snapshots](https://www.thundercompute.com/docs/mcp-server/operations/snapshots.md): Ask your AI agent to create, restore, or delete snapshots, or switch to another interface for direct control. - [Deleting Instances](https://www.thundercompute.com/docs/mcp-server/operations/deleting-instances.md): Ask your AI agent to delete Thunder instances, or switch to another interface for direct control. - [SSH on Thunder Compute](https://www.thundercompute.com/docs/mcp-server/operations/ssh.md): Ask your AI agent to work through Thunder over MCP, or switch to the console for manual SSH control. - [Jupyter Notebooks](https://www.thundercompute.com/docs/guides/running-jupyter-notebooks-on-thunder-compute.md): Execute Jupyter Notebooks on Thunder cloud GPUs. Configure remote kernels and process intensive data workloads in a notebook environment. - [Use Instance Templates for AI](https://www.thundercompute.com/docs/guides/using-instance-templates.md): Quickly deploy LLMs (Ollama) and AI image generators (ComfyUI) on Thunder using pre-configured instance templates. Get started fast. - [Speeding Up Snapshots](https://www.thundercompute.com/docs/guides/speeding-up-snapshots.md): Accelerate snapshot creation and restoration on Thunder. technical optimizations to reduce backup latency and improve data speed. - [Stopping Instances](https://www.thundercompute.com/docs/guides/stopping-instances.md): Manage instance states to optimize billing on Thunder. Learn how to pause and resume compute resources using snapshots. - [Using Docker](https://www.thundercompute.com/docs/guides/using-docker-on-thundercompute.md): Containerize applications using Docker on Thunder. Manage images, deploy containers, and optimize Docker environments on cloud GPU instances. - [Fine-tune with Unsloth Studio](https://www.thundercompute.com/docs/guides/unsloth-studio-finetune.md): Launch Unsloth Studio on a Thunder GPU instance and run a small Gemma 4 QLoRA fine-tune. - [Run Models](https://www.thundercompute.com/docs/guides/models.md): Choose a Thunder GPU, model format, and serving runtime for open model inference. - [Run Qwen3.6 27B](https://www.thundercompute.com/docs/guides/models/qwen3-6-27b.md): Launch Qwen3.6 27B dense on Thunder with a GPU-fitting quantization, tested llama.cpp commands, and OpenAI-compatible access. - [Run GPT‑OSS 120B on Thunder Compute](https://www.thundercompute.com/docs/guides/gpt-oss-running-locally-on-thunder-compute.md): Deploy GPT-OSS 120B on Thunder hardware. Initialize the large language model and configure the local environment for high-performance use. - [Self-host Deepseek R1](https://www.thundercompute.com/docs/guides/deepseek-r1-running-locally-on-thunder-compute.md): Self-host Deepseek R1 on Thunder cloud GPUs. Local model deployment and configure hardware for optimized inference performance. - [Thunder Compute Referral Program](https://www.thundercompute.com/docs/guides/referral-program.md): Earn credits by referring friends to Thunder. Get 3% of every dollar your referrals spend on GPU instances with our lifetime rewards program. - [Data Processing Addendum](https://www.thundercompute.com/docs/guides/data-processing-addendum.md): Review the Data Processing Addendum for Thunder. Audit legal terms, data handling protocols, and privacy compliance for your organization. - [Weights & Biases](https://www.thundercompute.com/docs/guides/weights-and-biases.md): Track, debug, and optimize GPU-heavy workloads on Thunder instances using Weights & Biases (wandb). - [MCP Server](https://www.thundercompute.com/docs/guides/mcp-server.md): Use Thunder with AI coding agents like Claude Code, Cursor, Devin Desktop, and Codex via the Model Context Protocol (MCP). - [Troubleshooting](https://www.thundercompute.com/docs/troubleshooting.md): Troubleshoot common Thunder errors. Find solutions for connection issues, function errors, SSH problems, and access logs. Get support via Discord. - [Billing](https://www.thundercompute.com/docs/billing.md): Understand Thunder's usage-based billing, payment methods, billing alerts, current rates, and tips for saving on GPU cloud costs. - [Restrictions](https://www.thundercompute.com/docs/restrictions.md): Analyze platform restrictions for Thunder services. Review technical limitations, usage policies, and resource constraints for accounts. - [Get credit alert](https://www.thundercompute.com/docs/api-reference/billing/get-credit-alert.md): Get the organization's credit-balance alert. The alert field is null when none is configured. - [Set credit alert](https://www.thundercompute.com/docs/api-reference/billing/set-credit-alert.md): Create or replace the organization's credit-balance alert. An email is sent to the organization's billing email when the credit balance drops below the threshold. - [Delete credit alert](https://www.thundercompute.com/docs/api-reference/billing/delete-credit-alert.md): Remove the organization's credit-balance alert - [Create instance](https://www.thundercompute.com/docs/api-reference/instances/create-instance.md): Create a new compute instance. - [List instances](https://www.thundercompute.com/docs/api-reference/instances/list-instances.md): Get a list of user's compute instances - [Add SSH key to instance](https://www.thundercompute.com/docs/api-reference/instances/add-ssh-key-to-instance.md): Append an SSH public key to an existing instance's `authorized_keys`. - [Delete instance](https://www.thundercompute.com/docs/api-reference/instances/delete-instance.md): Delete a compute instance by ID - [Modify instance](https://www.thundercompute.com/docs/api-reference/instances/modify-instance.md): Modify a running compute instance's resources - [Atomically update forwarded HTTP ports on an instance](https://www.thundercompute.com/docs/api-reference/instances/atomically-update-forwarded-http-ports-on-an-instance.md): Atomically add and remove forwarded HTTP ports on a running instance in a single transaction. Body: `{"add_ports": [8080], "remove_ports": [3000]}`. Adds are deduplicated and removing ports that are not currently forwarded is a no-op. Each forwarded port is reachable at https://-.thunder… - [Add an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/add-an-ssh-key.md): Add a new SSH public key to the authenticated user's organization - [List SSH keys](https://www.thundercompute.com/docs/api-reference/utilities/list-ssh-keys.md): Get a list of all SSH keys for the authenticated user's organization - [Delete an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/delete-an-ssh-key.md): Delete an SSH key by ID - [Get current pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-pricing.md): Retrieve current hourly pricing information for compute resources - [Get GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types, counts, and modes - [Get thunder templates](https://www.thundercompute.com/docs/api-reference/utilities/get-thunder-templates.md): Get available thunder templates for instance creation - [List API tokens](https://www.thundercompute.com/docs/api-reference/utilities/list-api-tokens.md): Returns the names of API tokens that belong to the authenticated user's organization. Token values/hashes are never returned. - [Get current v2 pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-v2-pricing.md): Retrieve current hourly pricing information for compute resources using mode-less spec keys - [Get v2 GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-v2-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types and counts using mode-less spec keys - [Get GPU availability](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-availability.md): Retrieve current GPU availability using mode-less spec keys - [Prepare a sandbox image from a Dockerfile](https://www.thundercompute.com/docs/api-reference/sandbox-images/prepare-a-sandbox-image-from-a-dockerfile.md) - [Import a sandbox image from a registry](https://www.thundercompute.com/docs/api-reference/sandbox-images/import-a-sandbox-image-from-a-registry.md) - [Get sandbox image status](https://www.thundercompute.com/docs/api-reference/sandbox-images/get-sandbox-image-status.md) - [List sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/list-sandboxes.md) - [Count daily sandbox creations](https://www.thundercompute.com/docs/api-reference/sandboxes/count-daily-sandbox-creations.md) - [List organization sandbox logs](https://www.thundercompute.com/docs/api-reference/sandboxes/list-organization-sandbox-logs.md) - [Issue an SSH certificate for the organization's sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/issue-an-ssh-certificate-for-the-organizations-sandboxes.md) - [Start a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/start-a-sandbox.md) - [Get a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/get-a-sandbox.md) - [List logs for a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/list-logs-for-a-sandbox.md) - [Update a sandbox network policy](https://www.thundercompute.com/docs/api-reference/sandboxes/update-a-sandbox-network-policy.md): Replaces the outbound network policy and waits for the owning node to apply it. - [Stop a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/stop-a-sandbox.md) - [Wait for a sandbox to become ready](https://www.thundercompute.com/docs/api-reference/sandboxes/wait-for-a-sandbox-to-become-ready.md) - [Create a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/create-a-snapshot.md): Create a new snapshot from a running instance - [List snapshots](https://www.thundercompute.com/docs/api-reference/snapshots/list-snapshots.md): Get a list of all snapshots for the authenticated user's organization - [Delete a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/delete-a-snapshot.md): Delete a snapshot by ID - [Quickstart](https://www.thundercompute.com/docs/vscode/quickstart.md): Configure Thunder using VSCode, Cursor, or Devin Desktop. Set up your development environment and establish remote connections with this guide. - [Technical Specifications](https://www.thundercompute.com/docs/technical-specs.md): Hardware specifications, networking details, and pre-installed software for Thunder instances - [Compatibility](https://www.thundercompute.com/docs/compatibility.md): Review known Thunder compatibility limitations, including managed memory, CUDA MPS, FFmpeg, and GPU profilers. - [Authentication](https://www.thundercompute.com/docs/vscode/operations/authentication.md): Authenticate Thunder services within VSCode. Manage API keys and SSH credentials to ensure secure access to your remote instances. - [Creating Instances](https://www.thundercompute.com/docs/vscode/operations/creating-instances.md): Provision new instances via the VSCode interface. Select hardware profiles and initialize remote environments on the Thunder platform. - [Connecting to Instances](https://www.thundercompute.com/docs/vscode/operations/connecting-to-instances.md): Connect to Thunder instances via VS Code. Configure SSH settings and manage remote development sessions for low-latency coding workflows. - [Monitoring Instances](https://www.thundercompute.com/docs/vscode/operations/monitoring-instances.md): Monitor instance performance and resource usage in VSCode. Track CPU, memory, and GPU metrics for your active Thunder cloud environment. - [File Transfers](https://www.thundercompute.com/docs/vscode/operations/file-transfers.md): Transfer files between local machines and Thunder instances. Manage data uploads and downloads within your VSCode development workflow. - [Port Forwarding](https://www.thundercompute.com/docs/vscode/operations/port-forwarding.md): Configure port forwarding to access local services on Thunder. Map internal ports to your local machine for efficient testing and debugging. - [Modifying Instances](https://www.thundercompute.com/docs/vscode/operations/modifying-instances.md): Modify instance configurations using VSCode. Update disk space and GPU counts for your active Thunder cloud resources. - [Snapshots](https://www.thundercompute.com/docs/vscode/operations/snapshots.md): Manage system snapshots directly within the VSCode environment. Create, restore, and organize backups to preserve your current development state. - [Deleting Instances](https://www.thundercompute.com/docs/vscode/operations/deleting-instances.md): Terminate and delete instances using VSCode management tools. Clean up organization hardware and release resources when tasks are complete. - [SSH on Thunder Compute](https://www.thundercompute.com/docs/vscode/operations/ssh.md): Manual SSH setup is only available in the Thunder Console. Use the console to manage saved keys and copy raw SSH connection details. - [Jupyter Notebooks](https://www.thundercompute.com/docs/guides/running-jupyter-notebooks-on-thunder-compute.md): Execute Jupyter Notebooks on Thunder cloud GPUs. Configure remote kernels and process intensive data workloads in a notebook environment. - [Use Instance Templates for AI](https://www.thundercompute.com/docs/guides/using-instance-templates.md): Quickly deploy LLMs (Ollama) and AI image generators (ComfyUI) on Thunder using pre-configured instance templates. Get started fast. - [Speeding Up Snapshots](https://www.thundercompute.com/docs/guides/speeding-up-snapshots.md): Accelerate snapshot creation and restoration on Thunder. technical optimizations to reduce backup latency and improve data speed. - [Stopping Instances](https://www.thundercompute.com/docs/guides/stopping-instances.md): Manage instance states to optimize billing on Thunder. Learn how to pause and resume compute resources using snapshots. - [Using Docker](https://www.thundercompute.com/docs/guides/using-docker-on-thundercompute.md): Containerize applications using Docker on Thunder. Manage images, deploy containers, and optimize Docker environments on cloud GPU instances. - [Fine-tune with Unsloth Studio](https://www.thundercompute.com/docs/guides/unsloth-studio-finetune.md): Launch Unsloth Studio on a Thunder GPU instance and run a small Gemma 4 QLoRA fine-tune. - [Run Models](https://www.thundercompute.com/docs/guides/models.md): Choose a Thunder GPU, model format, and serving runtime for open model inference. - [Run Qwen3.6 27B](https://www.thundercompute.com/docs/guides/models/qwen3-6-27b.md): Launch Qwen3.6 27B dense on Thunder with a GPU-fitting quantization, tested llama.cpp commands, and OpenAI-compatible access. - [Run GPT‑OSS 120B on Thunder Compute](https://www.thundercompute.com/docs/guides/gpt-oss-running-locally-on-thunder-compute.md): Deploy GPT-OSS 120B on Thunder hardware. Initialize the large language model and configure the local environment for high-performance use. - [Self-host Deepseek R1](https://www.thundercompute.com/docs/guides/deepseek-r1-running-locally-on-thunder-compute.md): Self-host Deepseek R1 on Thunder cloud GPUs. Local model deployment and configure hardware for optimized inference performance. - [Thunder Compute Referral Program](https://www.thundercompute.com/docs/guides/referral-program.md): Earn credits by referring friends to Thunder. Get 3% of every dollar your referrals spend on GPU instances with our lifetime rewards program. - [Data Processing Addendum](https://www.thundercompute.com/docs/guides/data-processing-addendum.md): Review the Data Processing Addendum for Thunder. Audit legal terms, data handling protocols, and privacy compliance for your organization. - [Weights & Biases](https://www.thundercompute.com/docs/guides/weights-and-biases.md): Track, debug, and optimize GPU-heavy workloads on Thunder instances using Weights & Biases (wandb). - [MCP Server](https://www.thundercompute.com/docs/guides/mcp-server.md): Use Thunder with AI coding agents like Claude Code, Cursor, Devin Desktop, and Codex via the Model Context Protocol (MCP). - [Thunder Compute Skill for OpenClaw](https://www.thundercompute.com/docs/guides/openclaw-thunder-compute-beta.md): Use Thunder from your own OpenClaw instance to see available GPUs, launch and stop instances, create snapshots, and report cost from chat. - [Troubleshooting](https://www.thundercompute.com/docs/troubleshooting.md): Troubleshoot common Thunder errors. Find solutions for connection issues, function errors, SSH problems, and access logs. Get support via Discord. - [Billing](https://www.thundercompute.com/docs/billing.md): Understand Thunder's usage-based billing, payment methods, billing alerts, current rates, and tips for saving on GPU cloud costs. - [Restrictions](https://www.thundercompute.com/docs/restrictions.md): Analyze platform restrictions for Thunder services. Review technical limitations, usage policies, and resource constraints for accounts. - [Get credit alert](https://www.thundercompute.com/docs/api-reference/billing/get-credit-alert.md): Get the organization's credit-balance alert. The alert field is null when none is configured. - [Set credit alert](https://www.thundercompute.com/docs/api-reference/billing/set-credit-alert.md): Create or replace the organization's credit-balance alert. An email is sent to the organization's billing email when the credit balance drops below the threshold. - [Delete credit alert](https://www.thundercompute.com/docs/api-reference/billing/delete-credit-alert.md): Remove the organization's credit-balance alert - [Create instance](https://www.thundercompute.com/docs/api-reference/instances/create-instance.md): Create a new compute instance. - [List instances](https://www.thundercompute.com/docs/api-reference/instances/list-instances.md): Get a list of user's compute instances - [Add SSH key to instance](https://www.thundercompute.com/docs/api-reference/instances/add-ssh-key-to-instance.md): Append an SSH public key to an existing instance's `authorized_keys`. - [Delete instance](https://www.thundercompute.com/docs/api-reference/instances/delete-instance.md): Delete a compute instance by ID - [Modify instance](https://www.thundercompute.com/docs/api-reference/instances/modify-instance.md): Modify a running compute instance's resources - [Atomically update forwarded HTTP ports on an instance](https://www.thundercompute.com/docs/api-reference/instances/atomically-update-forwarded-http-ports-on-an-instance.md): Atomically add and remove forwarded HTTP ports on a running instance in a single transaction. Body: `{"add_ports": [8080], "remove_ports": [3000]}`. Adds are deduplicated and removing ports that are not currently forwarded is a no-op. Each forwarded port is reachable at https://-.thunder… - [Add an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/add-an-ssh-key.md): Add a new SSH public key to the authenticated user's organization - [List SSH keys](https://www.thundercompute.com/docs/api-reference/utilities/list-ssh-keys.md): Get a list of all SSH keys for the authenticated user's organization - [Delete an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/delete-an-ssh-key.md): Delete an SSH key by ID - [Get current pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-pricing.md): Retrieve current hourly pricing information for compute resources - [Get GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types, counts, and modes - [Get thunder templates](https://www.thundercompute.com/docs/api-reference/utilities/get-thunder-templates.md): Get available thunder templates for instance creation - [List API tokens](https://www.thundercompute.com/docs/api-reference/utilities/list-api-tokens.md): Returns the names of API tokens that belong to the authenticated user's organization. Token values/hashes are never returned. - [Get current v2 pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-v2-pricing.md): Retrieve current hourly pricing information for compute resources using mode-less spec keys - [Get v2 GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-v2-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types and counts using mode-less spec keys - [Get GPU availability](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-availability.md): Retrieve current GPU availability using mode-less spec keys - [Prepare a sandbox image from a Dockerfile](https://www.thundercompute.com/docs/api-reference/sandbox-images/prepare-a-sandbox-image-from-a-dockerfile.md) - [Import a sandbox image from a registry](https://www.thundercompute.com/docs/api-reference/sandbox-images/import-a-sandbox-image-from-a-registry.md) - [Get sandbox image status](https://www.thundercompute.com/docs/api-reference/sandbox-images/get-sandbox-image-status.md) - [List sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/list-sandboxes.md) - [Count daily sandbox creations](https://www.thundercompute.com/docs/api-reference/sandboxes/count-daily-sandbox-creations.md) - [List organization sandbox logs](https://www.thundercompute.com/docs/api-reference/sandboxes/list-organization-sandbox-logs.md) - [Issue an SSH certificate for the organization's sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/issue-an-ssh-certificate-for-the-organizations-sandboxes.md) - [Start a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/start-a-sandbox.md) - [Get a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/get-a-sandbox.md) - [List logs for a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/list-logs-for-a-sandbox.md) - [Update a sandbox network policy](https://www.thundercompute.com/docs/api-reference/sandboxes/update-a-sandbox-network-policy.md): Replaces the outbound network policy and waits for the owning node to apply it. - [Stop a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/stop-a-sandbox.md) - [Wait for a sandbox to become ready](https://www.thundercompute.com/docs/api-reference/sandboxes/wait-for-a-sandbox-to-become-ready.md) - [Create a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/create-a-snapshot.md): Create a new snapshot from a running instance - [List snapshots](https://www.thundercompute.com/docs/api-reference/snapshots/list-snapshots.md): Get a list of all snapshots for the authenticated user's organization - [Delete a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/delete-a-snapshot.md): Delete a snapshot by ID - [Quickstart](https://www.thundercompute.com/docs/cli/quickstart.md): Configure the Thunder CLI for Mac, Windows or Linux. Deploy your first cloud instance in minutes. - [Technical Specifications](https://www.thundercompute.com/docs/technical-specs.md): Hardware specifications, networking details, and pre-installed software for Thunder instances - [Compatibility](https://www.thundercompute.com/docs/compatibility.md): Review known Thunder compatibility limitations, including managed memory, CUDA MPS, FFmpeg, and GPU profilers. - [Authentication](https://www.thundercompute.com/docs/cli/operations/authentication.md): Authenticate the Thunder CLI using API keys. Configure terminal credentials to manage cloud resources. - [Creating Instances](https://www.thundercompute.com/docs/cli/operations/creating-instances.md): Create new GPU instances via the CLI. Execute terminal commands to select hardware profiles and initialize remote environments. - [Connecting to Instances](https://www.thundercompute.com/docs/cli/operations/connecting-to-instances.md): Connect to Thunder instances using the CLI. Establish secure remote connections and manage terminal sessions for cloud GPU development. - [Monitoring Instances](https://www.thundercompute.com/docs/cli/operations/monitoring-instances.md): View active Thunder instances in the CLI. Audit status, IP, and GPU resources of your active cloud compute resources. - [File Transfers](https://www.thundercompute.com/docs/cli/operations/file-transfers.md): Move files to and from Thunder instances using the CLI. Upload or download data between local and cloud GPU hardware. - [Port Forwarding](https://www.thundercompute.com/docs/cli/operations/port-forwarding.md): Set up port forwarding using the CLI. Map remote instance ports to local endpoints for development, testing, and service access on cloud hardware. - [Modifying Instances](https://www.thundercompute.com/docs/cli/operations/modifying-instances.md): Reconfigure compute instances using the CLI. Swap GPUs, add vCPUs and RAM, or increase storage space for your compute resources. - [Snapshots](https://www.thundercompute.com/docs/cli/operations/snapshots.md): Create, delete, and restore instance snapshots using CLI commands. Ensure data persistence and safety. - [Deleting Instances](https://www.thundercompute.com/docs/cli/operations/deleting-instances.md): Delete instances using CLI commands. Release hardware resources and stop billing for active cloud GPUs. - [SSH on Thunder Compute](https://www.thundercompute.com/docs/cli/operations/ssh.md): Manual SSH setup is only available in the Thunder Console. Use the console to manage saved keys and copy raw SSH connection details. - [Jupyter Notebooks](https://www.thundercompute.com/docs/guides/running-jupyter-notebooks-on-thunder-compute.md): Execute Jupyter Notebooks on Thunder cloud GPUs. Configure remote kernels and process intensive data workloads in a notebook environment. - [Use Instance Templates for AI](https://www.thundercompute.com/docs/guides/using-instance-templates.md): Quickly deploy LLMs (Ollama) and AI image generators (ComfyUI) on Thunder using pre-configured instance templates. Get started fast. - [Speeding Up Snapshots](https://www.thundercompute.com/docs/guides/speeding-up-snapshots.md): Accelerate snapshot creation and restoration on Thunder. technical optimizations to reduce backup latency and improve data speed. - [Stopping Instances](https://www.thundercompute.com/docs/guides/stopping-instances.md): Manage instance states to optimize billing on Thunder. Learn how to pause and resume compute resources using snapshots. - [Using Docker](https://www.thundercompute.com/docs/guides/using-docker-on-thundercompute.md): Containerize applications using Docker on Thunder. Manage images, deploy containers, and optimize Docker environments on cloud GPU instances. - [Fine-tune with Unsloth Studio](https://www.thundercompute.com/docs/guides/unsloth-studio-finetune.md): Launch Unsloth Studio on a Thunder GPU instance and run a small Gemma 4 QLoRA fine-tune. - [Run Models](https://www.thundercompute.com/docs/guides/models.md): Choose a Thunder GPU, model format, and serving runtime for open model inference. - [Run Qwen3.6 27B](https://www.thundercompute.com/docs/guides/models/qwen3-6-27b.md): Launch Qwen3.6 27B dense on Thunder with a GPU-fitting quantization, tested llama.cpp commands, and OpenAI-compatible access. - [Run GPT‑OSS 120B on Thunder Compute](https://www.thundercompute.com/docs/guides/gpt-oss-running-locally-on-thunder-compute.md): Deploy GPT-OSS 120B on Thunder hardware. Initialize the large language model and configure the local environment for high-performance use. - [Self-host Deepseek R1](https://www.thundercompute.com/docs/guides/deepseek-r1-running-locally-on-thunder-compute.md): Self-host Deepseek R1 on Thunder cloud GPUs. Local model deployment and configure hardware for optimized inference performance. - [Thunder Compute Referral Program](https://www.thundercompute.com/docs/guides/referral-program.md): Earn credits by referring friends to Thunder. Get 3% of every dollar your referrals spend on GPU instances with our lifetime rewards program. - [Data Processing Addendum](https://www.thundercompute.com/docs/guides/data-processing-addendum.md): Review the Data Processing Addendum for Thunder. Audit legal terms, data handling protocols, and privacy compliance for your organization. - [Weights & Biases](https://www.thundercompute.com/docs/guides/weights-and-biases.md): Track, debug, and optimize GPU-heavy workloads on Thunder instances using Weights & Biases (wandb). - [MCP Server](https://www.thundercompute.com/docs/guides/mcp-server.md): Use Thunder with AI coding agents like Claude Code, Cursor, Devin Desktop, and Codex via the Model Context Protocol (MCP). - [Troubleshooting](https://www.thundercompute.com/docs/troubleshooting.md): Troubleshoot common Thunder errors. Find solutions for connection issues, function errors, SSH problems, and access logs. Get support via Discord. - [Billing](https://www.thundercompute.com/docs/billing.md): Understand Thunder's usage-based billing, payment methods, billing alerts, current rates, and tips for saving on GPU cloud costs. - [Restrictions](https://www.thundercompute.com/docs/restrictions.md): Analyze platform restrictions for Thunder services. Review technical limitations, usage policies, and resource constraints for accounts. - [Get credit alert](https://www.thundercompute.com/docs/api-reference/billing/get-credit-alert.md): Get the organization's credit-balance alert. The alert field is null when none is configured. - [Set credit alert](https://www.thundercompute.com/docs/api-reference/billing/set-credit-alert.md): Create or replace the organization's credit-balance alert. An email is sent to the organization's billing email when the credit balance drops below the threshold. - [Delete credit alert](https://www.thundercompute.com/docs/api-reference/billing/delete-credit-alert.md): Remove the organization's credit-balance alert - [Create instance](https://www.thundercompute.com/docs/api-reference/instances/create-instance.md): Create a new compute instance. - [List instances](https://www.thundercompute.com/docs/api-reference/instances/list-instances.md): Get a list of user's compute instances - [Add SSH key to instance](https://www.thundercompute.com/docs/api-reference/instances/add-ssh-key-to-instance.md): Append an SSH public key to an existing instance's `authorized_keys`. - [Delete instance](https://www.thundercompute.com/docs/api-reference/instances/delete-instance.md): Delete a compute instance by ID - [Modify instance](https://www.thundercompute.com/docs/api-reference/instances/modify-instance.md): Modify a running compute instance's resources - [Atomically update forwarded HTTP ports on an instance](https://www.thundercompute.com/docs/api-reference/instances/atomically-update-forwarded-http-ports-on-an-instance.md): Atomically add and remove forwarded HTTP ports on a running instance in a single transaction. Body: `{"add_ports": [8080], "remove_ports": [3000]}`. Adds are deduplicated and removing ports that are not currently forwarded is a no-op. Each forwarded port is reachable at https://-.thunder… - [Add an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/add-an-ssh-key.md): Add a new SSH public key to the authenticated user's organization - [List SSH keys](https://www.thundercompute.com/docs/api-reference/utilities/list-ssh-keys.md): Get a list of all SSH keys for the authenticated user's organization - [Delete an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/delete-an-ssh-key.md): Delete an SSH key by ID - [Get current pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-pricing.md): Retrieve current hourly pricing information for compute resources - [Get GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types, counts, and modes - [Get thunder templates](https://www.thundercompute.com/docs/api-reference/utilities/get-thunder-templates.md): Get available thunder templates for instance creation - [List API tokens](https://www.thundercompute.com/docs/api-reference/utilities/list-api-tokens.md): Returns the names of API tokens that belong to the authenticated user's organization. Token values/hashes are never returned. - [Get current v2 pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-v2-pricing.md): Retrieve current hourly pricing information for compute resources using mode-less spec keys - [Get v2 GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-v2-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types and counts using mode-less spec keys - [Get GPU availability](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-availability.md): Retrieve current GPU availability using mode-less spec keys - [Prepare a sandbox image from a Dockerfile](https://www.thundercompute.com/docs/api-reference/sandbox-images/prepare-a-sandbox-image-from-a-dockerfile.md) - [Import a sandbox image from a registry](https://www.thundercompute.com/docs/api-reference/sandbox-images/import-a-sandbox-image-from-a-registry.md) - [Get sandbox image status](https://www.thundercompute.com/docs/api-reference/sandbox-images/get-sandbox-image-status.md) - [List sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/list-sandboxes.md) - [Count daily sandbox creations](https://www.thundercompute.com/docs/api-reference/sandboxes/count-daily-sandbox-creations.md) - [List organization sandbox logs](https://www.thundercompute.com/docs/api-reference/sandboxes/list-organization-sandbox-logs.md) - [Issue an SSH certificate for the organization's sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/issue-an-ssh-certificate-for-the-organizations-sandboxes.md) - [Start a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/start-a-sandbox.md) - [Get a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/get-a-sandbox.md) - [List logs for a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/list-logs-for-a-sandbox.md) - [Update a sandbox network policy](https://www.thundercompute.com/docs/api-reference/sandboxes/update-a-sandbox-network-policy.md): Replaces the outbound network policy and waits for the owning node to apply it. - [Stop a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/stop-a-sandbox.md) - [Wait for a sandbox to become ready](https://www.thundercompute.com/docs/api-reference/sandboxes/wait-for-a-sandbox-to-become-ready.md) - [Create a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/create-a-snapshot.md): Create a new snapshot from a running instance - [List snapshots](https://www.thundercompute.com/docs/api-reference/snapshots/list-snapshots.md): Get a list of all snapshots for the authenticated user's organization - [Delete a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/delete-a-snapshot.md): Delete a snapshot by ID - [Thunder GPU Cloud Quickstart](https://www.thundercompute.com/docs/console/quickstart.md): Launch your first on-demand cloud GPU instance with Thunder. - [Technical Specifications](https://www.thundercompute.com/docs/technical-specs.md): Hardware specifications, networking details, and pre-installed software for Thunder instances - [Compatibility](https://www.thundercompute.com/docs/compatibility.md): Review known Thunder compatibility limitations, including managed memory, CUDA MPS, FFmpeg, and GPU profilers. - [Authentication](https://www.thundercompute.com/docs/console/operations/authentication.md): Secure your Thunder Console account. Manage user permissions, API keys, and authentication credentials for all yourcompute resources. - [Creating Instances](https://www.thundercompute.com/docs/console/operations/creating-instances.md): Create new GPU instances via the Cloud Console. Select hardware profiles and initialize environments on the Thunder platform. - [Connecting to Instances](https://www.thundercompute.com/docs/console/operations/connecting-to-instances.md): Connect to Thunder instances created in the Console. Use VSCode or the CLI to manage active cloud GPU. - [Monitoring Instances](https://www.thundercompute.com/docs/console/operations/monitoring-instances.md): View active instances in the Thunder Console. Audit status, IP, and GPU resources of your active cloud compute resources. - [File Transfers](https://www.thundercompute.com/docs/console/operations/file-transfers.md): File transfers cannot be done using the Thunder Console. Use the CLI or VSCode to move data between local machines and cloud GPU hardware. - [Port Forwarding](https://www.thundercompute.com/docs/console/operations/port-forwarding.md): Set up port forwarding using the Thunder Console. Map remote instance ports to local endpoints for development, testing, and service access. - [Modifying Instances](https://www.thundercompute.com/docs/console/operations/modifying-instances.md): Update instance configurations in the Thunder Console. Swap GPUs, add vCPUs and RAM, or increase storage space for your compute resources. - [Snapshots](https://www.thundercompute.com/docs/console/operations/snapshots.md): Create, delete, and restore instance snapshots using the Thunder Console. Ensure data persistence and safety. - [Deleting Instances](https://www.thundercompute.com/docs/console/operations/deleting-instances.md): Delete instances using the Thunder Console. Release hardware resources and stop billing for active cloud GPUs. - [SSH on Thunder Compute](https://www.thundercompute.com/docs/console/operations/ssh.md): Use the Thunder Console to manage SSH keys and copy the manual SSH command for an instance. - [Jupyter Notebooks](https://www.thundercompute.com/docs/guides/running-jupyter-notebooks-on-thunder-compute.md): Execute Jupyter Notebooks on Thunder cloud GPUs. Configure remote kernels and process intensive data workloads in a notebook environment. - [Use Instance Templates for AI](https://www.thundercompute.com/docs/guides/using-instance-templates.md): Quickly deploy LLMs (Ollama) and AI image generators (ComfyUI) on Thunder using pre-configured instance templates. Get started fast. - [Speeding Up Snapshots](https://www.thundercompute.com/docs/guides/speeding-up-snapshots.md): Accelerate snapshot creation and restoration on Thunder. technical optimizations to reduce backup latency and improve data speed. - [Stopping Instances](https://www.thundercompute.com/docs/guides/stopping-instances.md): Manage instance states to optimize billing on Thunder. Learn how to pause and resume compute resources using snapshots. - [Using Docker](https://www.thundercompute.com/docs/guides/using-docker-on-thundercompute.md): Containerize applications using Docker on Thunder. Manage images, deploy containers, and optimize Docker environments on cloud GPU instances. - [Fine-tune with Unsloth Studio](https://www.thundercompute.com/docs/guides/unsloth-studio-finetune.md): Launch Unsloth Studio on a Thunder GPU instance and run a small Gemma 4 QLoRA fine-tune. - [Run Models](https://www.thundercompute.com/docs/guides/models.md): Choose a Thunder GPU, model format, and serving runtime for open model inference. - [Run Qwen3.6 27B](https://www.thundercompute.com/docs/guides/models/qwen3-6-27b.md): Launch Qwen3.6 27B dense on Thunder with a GPU-fitting quantization, tested llama.cpp commands, and OpenAI-compatible access. - [Run GPT‑OSS 120B on Thunder Compute](https://www.thundercompute.com/docs/guides/gpt-oss-running-locally-on-thunder-compute.md): Deploy GPT-OSS 120B on Thunder hardware. Initialize the large language model and configure the local environment for high-performance use. - [Self-host Deepseek R1](https://www.thundercompute.com/docs/guides/deepseek-r1-running-locally-on-thunder-compute.md): Self-host Deepseek R1 on Thunder cloud GPUs. Local model deployment and configure hardware for optimized inference performance. - [Thunder Compute Referral Program](https://www.thundercompute.com/docs/guides/referral-program.md): Earn credits by referring friends to Thunder. Get 3% of every dollar your referrals spend on GPU instances with our lifetime rewards program. - [Data Processing Addendum](https://www.thundercompute.com/docs/guides/data-processing-addendum.md): Review the Data Processing Addendum for Thunder. Audit legal terms, data handling protocols, and privacy compliance for your organization. - [Weights & Biases](https://www.thundercompute.com/docs/guides/weights-and-biases.md): Track, debug, and optimize GPU-heavy workloads on Thunder instances using Weights & Biases (wandb). - [MCP Server](https://www.thundercompute.com/docs/guides/mcp-server.md): Use Thunder with AI coding agents like Claude Code, Cursor, Devin Desktop, and Codex via the Model Context Protocol (MCP). - [Troubleshooting](https://www.thundercompute.com/docs/troubleshooting.md): Troubleshoot common Thunder errors. Find solutions for connection issues, function errors, SSH problems, and access logs. Get support via Discord. - [Billing](https://www.thundercompute.com/docs/billing.md): Understand Thunder's usage-based billing, payment methods, billing alerts, current rates, and tips for saving on GPU cloud costs. - [Restrictions](https://www.thundercompute.com/docs/restrictions.md): Analyze platform restrictions for Thunder services. Review technical limitations, usage policies, and resource constraints for accounts. - [Get credit alert](https://www.thundercompute.com/docs/api-reference/billing/get-credit-alert.md): Get the organization's credit-balance alert. The alert field is null when none is configured. - [Set credit alert](https://www.thundercompute.com/docs/api-reference/billing/set-credit-alert.md): Create or replace the organization's credit-balance alert. An email is sent to the organization's billing email when the credit balance drops below the threshold. - [Delete credit alert](https://www.thundercompute.com/docs/api-reference/billing/delete-credit-alert.md): Remove the organization's credit-balance alert - [Create instance](https://www.thundercompute.com/docs/api-reference/instances/create-instance.md): Create a new compute instance. - [List instances](https://www.thundercompute.com/docs/api-reference/instances/list-instances.md): Get a list of user's compute instances - [Add SSH key to instance](https://www.thundercompute.com/docs/api-reference/instances/add-ssh-key-to-instance.md): Append an SSH public key to an existing instance's `authorized_keys`. - [Delete instance](https://www.thundercompute.com/docs/api-reference/instances/delete-instance.md): Delete a compute instance by ID - [Modify instance](https://www.thundercompute.com/docs/api-reference/instances/modify-instance.md): Modify a running compute instance's resources - [Atomically update forwarded HTTP ports on an instance](https://www.thundercompute.com/docs/api-reference/instances/atomically-update-forwarded-http-ports-on-an-instance.md): Atomically add and remove forwarded HTTP ports on a running instance in a single transaction. Body: `{"add_ports": [8080], "remove_ports": [3000]}`. Adds are deduplicated and removing ports that are not currently forwarded is a no-op. Each forwarded port is reachable at https://-.thunder… - [Add an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/add-an-ssh-key.md): Add a new SSH public key to the authenticated user's organization - [List SSH keys](https://www.thundercompute.com/docs/api-reference/utilities/list-ssh-keys.md): Get a list of all SSH keys for the authenticated user's organization - [Delete an SSH key](https://www.thundercompute.com/docs/api-reference/utilities/delete-an-ssh-key.md): Delete an SSH key by ID - [Get current pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-pricing.md): Retrieve current hourly pricing information for compute resources - [Get GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types, counts, and modes - [Get thunder templates](https://www.thundercompute.com/docs/api-reference/utilities/get-thunder-templates.md): Get available thunder templates for instance creation - [List API tokens](https://www.thundercompute.com/docs/api-reference/utilities/list-api-tokens.md): Returns the names of API tokens that belong to the authenticated user's organization. Token values/hashes are never returned. - [Get current v2 pricing](https://www.thundercompute.com/docs/api-reference/utilities/get-current-v2-pricing.md): Retrieve current hourly pricing information for compute resources using mode-less spec keys - [Get v2 GPU specifications](https://www.thundercompute.com/docs/api-reference/utilities/get-v2-gpu-specifications.md): Retrieve GPU spec configurations for all supported GPU types and counts using mode-less spec keys - [Get GPU availability](https://www.thundercompute.com/docs/api-reference/utilities/get-gpu-availability.md): Retrieve current GPU availability using mode-less spec keys - [Prepare a sandbox image from a Dockerfile](https://www.thundercompute.com/docs/api-reference/sandbox-images/prepare-a-sandbox-image-from-a-dockerfile.md) - [Import a sandbox image from a registry](https://www.thundercompute.com/docs/api-reference/sandbox-images/import-a-sandbox-image-from-a-registry.md) - [Get sandbox image status](https://www.thundercompute.com/docs/api-reference/sandbox-images/get-sandbox-image-status.md) - [List sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/list-sandboxes.md) - [Count daily sandbox creations](https://www.thundercompute.com/docs/api-reference/sandboxes/count-daily-sandbox-creations.md) - [List organization sandbox logs](https://www.thundercompute.com/docs/api-reference/sandboxes/list-organization-sandbox-logs.md) - [Issue an SSH certificate for the organization's sandboxes](https://www.thundercompute.com/docs/api-reference/sandboxes/issue-an-ssh-certificate-for-the-organizations-sandboxes.md) - [Start a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/start-a-sandbox.md) - [Get a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/get-a-sandbox.md) - [List logs for a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/list-logs-for-a-sandbox.md) - [Update a sandbox network policy](https://www.thundercompute.com/docs/api-reference/sandboxes/update-a-sandbox-network-policy.md): Replaces the outbound network policy and waits for the owning node to apply it. - [Stop a sandbox](https://www.thundercompute.com/docs/api-reference/sandboxes/stop-a-sandbox.md) - [Wait for a sandbox to become ready](https://www.thundercompute.com/docs/api-reference/sandboxes/wait-for-a-sandbox-to-become-ready.md) - [Create a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/create-a-snapshot.md): Create a new snapshot from a running instance - [List snapshots](https://www.thundercompute.com/docs/api-reference/snapshots/list-snapshots.md): Get a list of all snapshots for the authenticated user's organization - [Delete a snapshot](https://www.thundercompute.com/docs/api-reference/snapshots/delete-a-snapshot.md): Delete a snapshot by ID ## OpenAPI Specs - [openapi](https://api.thundercompute.com:8443/openapi.json) - [package](/docs/package.json) - [swagger](/docs/swagger.json) ## Optional - [Support](https://www.thundercompute.com/contact)