DOCSEARCHbyAlgolia
Submit Docs

Runpod Documentation

website·Indexed 40 days ago

https://docs.runpod.io

Documentation for Runpod's GPU cloud platform, including guides on billing, account management, and the REST API v2 for managing Pods, Serverless endpoints, Clusters, and network volumes.

Tokens
194.7K
Snippets
969
Records
1.5K
Agent score
70%

What's inside Runpod

  1. Overview of RunPod Community Solutions

    RunPod features a collection of community-created tools and workflows developed by users to solve common challenges. These tools are not officially maintained by RunPod and are provided as-is. Users are advised to test these solutions in a development environment, review available source code, and use them at their own discretion.

  2. Overview of Flash application management

    A Flash app is the top-level container that groups deployment environments, build artifacts, and configuration. You can manage these apps using the flash app command group.
    flash app <subcommand> [OPTIONS]

  3. Overview of Runpod Flash

    Runpod Flash is a Python SDK that allows developers to build autoscaling AI/ML applications using local code. By using the @Endpoint decorator, developers can define hardware requirements (GPU/CPU), remote functions, and dependencies locally. Flash automatically handles GPU/CPU provisioning and worker scaling on Runpod Serverless.

  4. Overview of RunPod Global Networking

    Global networking creates a secure, private network that connects all Pods within a single RunPod account, regardless of their physical data center location. It provides each Pod with a private IP address and DNS name, allowing Pod-to-Pod communication as if they were on the same local network. The network operates at 100 Mbps and is currently only available for NVIDIA GPU Pods.

  5. Overview of Runpod Public Endpoints

    Runpod Public Endpoints provide instant API access to pre-deployed AI models for image, video, audio, and text generation. Users can make requests to these models using an API key without needing to manage their own deployment or GPU infrastructure.

  6. Overview of Runpod Cost Centers

    Cost centers allow you to attach billing labels to Runpod resources to track and manage spending across an organization. They enable attribution of charges to specific teams, projects, or departments, helping users reconcile GPU and storage spend with internal budgets.

  7. Overview of Runpod Slurm Clusters

    Runpod Slurm Clusters provide a managed high-performance computing (HPC) and scheduling solution. They feature zero-configuration setup where Slurm and munge are pre-installed, instant provisioning, and automatic role assignment for controller and agent nodes.

  8. Overview of RunPod Serverless Batch jobs

    Batch jobs allow you to submit large collections of Serverless inference requests as named batches. They run asynchronously on dedicated workers, ensuring that bulk workloads do not delay interactive requests sent via the standard /run endpoint. Batch jobs are ideal for offline workloads that can tolerate multi-hour latency, such as nightly dataset processing, pre-computing embeddings, or running evaluations.

  9. Overview of RunPod Instant Clusters

    RunPod Instant Clusters provide fully managed multi-node compute clusters designed for distributed workloads such as large-scale AI inference and multi-node training. They eliminate the need to manually manage infrastructure, networking, or cluster configuration by providing pre-configured static IPs and environment variables. Key features include high-speed inter-node networking (1600-3200 Gbps) for efficient gradient synchronization and data movement, and on-demand deployment.

  10. Overview of Moonshot Kimi on Runpod

    Moonshot Kimi is a family of models designed for advanced reasoning, chat, and coding. These models support extended thinking with visible reasoning traces. They are available via Runpod Public Endpoints and are fully compatible with the OpenAI API.

  11. Overview of Runpod secrets

    Runpod secrets are encrypted strings used to store sensitive information (API keys, passwords, authentication tokens) separately from Pod template configurations. They provide security by ensuring sensitive data is never displayed in plain text after creation, allow for reusability across multiple templates and Pods, and are tied to specific accounts or teams for access control.

  12. Overview of Runpod MCP Servers

    Runpod provides two Model Context Protocol (MCP) servers to connect AI tools and coding agents to their ecosystem:

    1. Runpod API MCP server: Allows management of Pods, endpoints, templates, volumes, and registries via the Runpod REST API. Requires authentication (OAuth or API key).
    2. Runpod docs MCP server: Provides searchable access to Runpod documentation, features, and code examples. No authentication is required.