Cloudflare Computer: Give Your Agent a Virtual Filesystem

Cloudflare Computer: Give Your Agent a Virtual Filesystem

Summary

Cloudflare Computer is an innovative project that provides a virtual filesystem within a Durable Object, allowing agents to interact with a persistent state. It offers pluggable execution backends, including containerized Linux environments and serverless Workers, enabling flexible and powerful agent computing. Currently in preview, it's ideal for experiments and prototypes leveraging Cloudflare's edge infrastructure.

At a glance

Added to OSRepos
August 14, 2026
Last analyzed
August 14, 2026
View on GitHub

Use at your own risk

OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.

Introduction

Cloudflare Computer is an innovative project designed to provide a virtual filesystem within a Cloudflare Durable Object. This system maintains its authoritative state in SQLite, offering a persistent and reliable storage solution. A key feature is its pluggable execution surface, workspace.runtime, which allows for diverse computing environments.

It currently ships with three distinct backends:

  • Container: Projects the SQLite state into a sandbox container as a real FUSE mount, providing a full Linux userland with real binaries and network access.
  • Isolate shell: Runs just-bash in a Dynamic Worker, interacting with the authoritative Workspace over Workers RPC.
  • Isolate JavaScript: Executes an ECMAScript module in a fresh Dynamic Worker, complete with structured input/results and Workspace-backed filesystem capabilities.

Please note, Cloudflare Computer is currently in PREVIEW ONLY. Its APIs are unstable, and the design is subject to change. It is suitable for experiments, exploration, and prototypes, but not yet for production use.

Installation

To begin building with Cloudflare Computer, you can install the core package:

npm install @cloudflare/computer

For detailed installation steps, entrypoint maps, and worked examples of the fs and runtime surfaces, please refer to the README of the @cloudflare/computer package within the repository.

Examples

The repository includes an examples/ directory, showcasing various ways to utilize Cloudflare Computer. Each example is a Worker workspace with its own README:

  • Container: Demonstrates running computerd inside a container, mounting a workspace, and communicating with a Durable Object over capnweb.
  • Worker Shell: Shows a similar HTTP surface to the container example, but the shell runs just-bash in a Dynamic Worker.
  • Worker JavaScript: Mirrors the worker-shell example, but executes an ECMAScript module in a Dynamic Worker instead of a shell command.
  • Think: An example of a @cloudflare/think chat agent that uses the workspace as its working directory.
  • Tutorial: A step-by-step guide building an agent that writes a markdown recipe card and converts it to a PDF using pandoc in a container.
  • Artifacts: Generates a Worker project within a workspace and publishes it to Cloudflare Artifacts.
  • Assets: Turns a prompt into an image using Workers AI, writes it to the workspace, and returns a shareable link.

Why Use It

Cloudflare Computer offers a unique platform for developing intelligent agents and serverless applications that require persistent state and flexible execution environments. By combining a virtual filesystem with Cloudflare's Durable Objects and Workers, it enables developers to:

  • Provide agents with a "working directory" that persists across invocations.
  • Execute code in diverse environments, from full Linux containers to lightweight JavaScript or shell Workers.
  • Leverage Cloudflare's global network for low-latency, scalable operations.
  • Experiment with advanced agent architectures and serverless computing patterns.

Source repository

Open the original repository on GitHub.

17 counted GitHub visits

View on GitHub

Related repositories

Similar repositories that may be relevant next.

OrcaReplay: Time Travel for AI Agents, Debugging and Evaluation

OrcaReplay: Time Travel for AI Agents, Debugging and Evaluation

October 2, 2026

OrcaReplay introduces "time travel" capabilities for AI agents, allowing developers to record, replay, fork, and debug any agent run with any model. It addresses the challenges of AI agent debugging by providing byte-for-byte reproducibility, offline analysis, and the ability to compare different models from specific checkpoints. This tool, built by the OrcaRouter.ai team, enhances observability and control over complex agent behaviors.

Agent DebuggingAI AgentsLLM Agents
OpenMake LLM: Self-Hosted AI Workspace for Local and Open-Weight LLMs

OpenMake LLM: Self-Hosted AI Workspace for Local and Open-Weight LLMs

October 1, 2026

OpenMake LLM is an open-source, self-hosted AI workspace for local and open-weight LLMs. It coordinates specialized models, autonomous agents, and tools for deep research and artifact generation. This platform supports vLLM, LiteLLM, and BYOK providers, offering a robust environment for managing AI workloads.

AI AgentsAI WorkspaceSelf Hosted AI
ZenNotes: Keyboard-First Markdown Notes with Vim, Diagrams, and MCP Integration

ZenNotes: Keyboard-First Markdown Notes with Vim, Diagrams, and MCP Integration

October 1, 2026

ZenNotes is a versatile, keyboard-first Markdown notes app designed for speed and flexibility. It stores notes as plain Markdown files, offering Vim-friendly editing, diagram support, and integration with MCP tools. Available as a desktop app (Electron) and a self-hosted web app, ZenNotes provides a powerful solution for organizing your thoughts.

ElectronLocal FirstMarkdown
lat.md: A Knowledge Graph for Your Codebase, Written in Markdown

lat.md: A Knowledge Graph for Your Codebase, Written in Markdown

September 26, 2026

lat.md is an innovative tool that transforms your codebase knowledge into an interconnected graph of markdown files. It helps both AI agents and human developers quickly understand project architecture, business logic, and design decisions. By integrating directly into your project, lat.md ensures documentation remains consistent and up-to-date.

TypeScriptDocumentationKnowledge Graph
OS
OSRepos

Analysis and discovery of open source repositories. Find interesting projects and follow their updates.

Monitor your website with YourWebsiteScore

OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of third-party repository code is at your own risk. Always review source code, dependencies, licenses, and security implications before running anything.

© 2025 OSRepos. Built with Nuxt 3 and lots of ❤️