Octomind Cloud + Hub

Product Introduction

1. Definition: Octomind Cloud + Hub is a cloud-based AI agent runtime and orchestration platform. Technically, it is a managed container service pre-configured with the open-source Octomind agent runtime, providing a seamless interface (web panel, CLI, API) to deploy, interact with, and manage persistent AI agent sessions.
2. Core Value Proposition: It exists to eliminate the friction of running AI agents locally by providing a serverless-like, persistent, and device-agnostic environment. Its primary value is enabling long-running AI agent tasks that survive laptop closures, with per-second billing, zero model API key management, and session persistence across devices.

Main Features

1. Persistent Cloud Machines: Users provision on-demand containers ("Machines") with selectable CPU/RAM profiles and pre-built images (Base, Node, Python, Rust). These containers run the Octomind binary and persist user sessions and files. They automatically suspend after configurable idle periods, switching to a low-cost storage-only billing model. The technology stack is based on container orchestration, with a custom runtime built in Rust (Apache 2.0).
2. Unified Model Gateway & Hub: The platform includes a built-in model gateway, providing direct access to 21 AI models without requiring users to manage individual API keys from providers like OpenAI or Anthropic. The accompanying Octomind Hub is a central catalog for discovering, comparing, and selecting these models based on transparent per-token pricing and benchmark performance, usable both in the cloud and with the local CLI.
3. Durable Sessions & Shared Filesystem: Every interaction with an agent is a "Session" that is fully recorded and replayable from any device. A shared directory is mounted across all a user's Machines, providing a persistent workspace for code indexes, agent memory, and project files. This creates a continuous context that follows the user, unlike ephemeral local processes.
4. Full Programmatic Control (Developer API): Every action in the web panel is mirrored by a comprehensive REST API, available on all plans including Free. Developers can programmatically create machines, initiate agent sessions, stream responses, and manage files using idempotent POST requests and Server-Sent Events (SSE), enabling integration into CI/CD pipelines, cron jobs, or custom applications.

Problems Solved

1. Pain Point: Agent Session Fragility. Traditional local AI agent processes terminate when the initiating laptop sleeps, disconnects, or is closed, killing long-running tasks like code refactors, research sweeps, or batch processing.
2. Target Audience: Software Developers & Engineers who use AI coding assistants for complex, multi-step tasks; Researchers & Data Scientists conducting extended AI-driven analysis; DevOps Engineers seeking to automate agent workloads in the cloud.
3. Use Cases: Asynchronous Code Refactoring: Start a major codebase refactor from a desktop, review and guide the agent's plan from a mobile device later, and merge the final PR. Heavy Computational Offloading: Index a massive code repository or run a full test suite on a cloud Machine to conserve local laptop battery and CPU. Persistent AI-Assisted Workflows: Run the built-in multi-step "research, write, audit, deliver" pipelines without worrying about process interruption.

Unique Advantages

1. Differentiation: Unlike pure local agent frameworks (e.g., Cline, Aider) or generic cloud VMs, Octomind Cloud provides a purpose-built, agent-optimized environment with persistent state. Unlike other cloud AI platforms, it offers true per-second granular billing with automatic suspension and a transparent, open-source runtime, avoiding vendor lock-in and unpredictable costs.
2. Key Innovation: The "session as a durable primitive" model. The platform treats an AI agent interaction not as a transient API call but as a stateful, replayable, and device-independent session. This is coupled with the integrated model Hub, which abstracts away the complexity of procuring and managing multiple AI model APIs, offering a unified, billed-through interface.

Frequently Asked Questions (FAQ)

1. How much does Octomind Cloud cost when I'm not using it? When idle for a short period (5 minutes on Free, 15 on Pro), Machines suspend automatically. While suspended, you pay only for storage ($0.10/GB-month beyond the included 10GB). While idle but not suspended, a minimal memory-only rate applies (~$0.01/hr for a Small machine).
2. Can I run my own AI models or bring my own API keys to Octomind? The Octomind Cloud platform uses its own unified model gateway and does not support injecting personal API keys. For full control over models, you can use the open-source Octomind runtime locally with the CLI, connecting it to your preferred model providers directly.
3. What happens to my data and work if I cancel my Octomind Pro subscription? Upon cancellation, your account, Machines, sessions, and files are preserved and not deleted for a 30-day grace period. You can resubscribe within this window and resume exactly where you left off. You can also download all your files from the panel at any time.
4. Is there a terminal/SSH access to the Octomind Cloud Machines? Yes. Beyond the chat interface, an "Advanced mode" provides a full web-based terminal (a real bash shell) into the container. This allows direct command-line execution of the octomind binary, installation of additional system packages, and direct file manipulation.
5. How does Octomind Cloud compare to running agents on services like Replit or GitHub Codespaces? While those are general-purpose cloud development environments, Octomind Cloud is specialized for AI agent orchestration. Key differences include: built-in AI model gateway, native agent session persistence and replay, per-second billing focused on agent runtime, and workflows designed specifically for multi-step AI tasks like code generation and auditing.