Practical overview
When it fits
Use Sandboxd when you want the app-building runtime and previews to run on infrastructure you operate, or when you need a console or scriptable API for isolated app sandboxes. The live demo is useful for exploring the sample console without installing the service.
Who it is for
- Developers who want to build and preview agent-generated apps on their own server.
- Teams that need isolated environments for prototyping or reviewing generated applications.
- Operators who want to provision and control app sandboxes through scripts or API calls.
- Users who want to run curated self-hosted applications alongside generated projects.
Documented uses
- Prompt a coding agent to build an app in an isolated container and open its preview URL.
- Create and inspect apps through a browser console, edit files, and review a Git diff before pushing.
- Import a public Git repository and have a coding agent work on it.
- Launch curated applications or start from React/Vite, Next.js, and FastAPI scaffolds.
- Use API calls to create sandboxes and submit agent tasks.
Requirements & limits
The self-hosted project is open source under the MIT license. The repository documents a live browser demo with sample data and no installation. Self-hosting requires Docker with the Compose plugin and Git on Linux; macOS via Docker Desktop is best-effort.
- The documented service stack uses Go, Docker, Traefik for preview URLs, and SQLite for state.
- The console is optional; the engine can run headlessly and actions are available through a scriptable API.
- The quick start documents amd64 and arm64 support.
- Public Git repositories can be imported without credentials, according to the project documentation.
- The self-hosted setup is distinct from the browser demo, which uses sample data.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Sandboxd and open its console
Start the self-hosted stack with the web console and retrieve the console login.
Before you start
- Linux server or host with Docker and the Compose plugin
- Git
- macOS via Docker Desktop is described as best-effort
Steps
Run the documented install script: the command below.
curl -fsSL https://raw.githubusercontent.com/tastyeffectco/sandboxd/main/install.sh | bashWait for the installer to build images and start the stack; it prints the console URL and generated login.
Open the printed console URL, connect an agent under Settings, then create an app.
Expected outcome
The self-hosted stack starts with its browser console, ready for agent configuration and app creation.
Guide limitations
- The guide documents setup and console access, not a completed app build.
- macOS via Docker Desktop is best-effort.
Create a sandbox and submit an agent task through the API
Use the documented API sequence to create a sandbox and send it a coding prompt.
Before you start
- A running Sandboxd instance; the documented local API address is
http://127.0.0.1:9090. - An agent API key to connect an agent.
- Docker and the Compose plugin, plus Git, for the documented self-hosted setup.
Steps
Set the
APIshell variable to the address of the running Sandboxd API.Connect an agent by posting its API key to
/v1/agents/claude-code/api-key.Create a sandbox with port 3000 and capture the returned sandbox ID.
Post a task with a prompt and agent name to
/v1/sandboxes/$ID/tasks.Open the documented preview URL pattern
http://s-$ID-3000.preview.localhost.
Expected outcome
The API sequence creates a sandbox, submits a coding task, and identifies the preview URL pattern for the app.
Guide limitations
- The documented example uses a local API address and a task targeting port 3000.
- The guide describes the request sequence; it does not verify that a task completed successfully.
Compare options for AI coding assistance
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Sandboxd | Prompt a coding agent to build an app in an isolated container and open its preview URL. | Docker, Self-hosted, Server | The self-hosted project is open source under the MIT license. The repository documents a live browser demo with sample data and no installation. Self-hosting requires Docker with the Compose plugin and Git on Linux; macOS via Docker Desktop is best-effort. |
| Nasiko | Detect supported coding-agent CLIs and inspect their versions and configuration locations. | CLI | The documented discovery command requires no account. Reporting and routing require an active control plane and valid login. The Docker Compose path requires Docker Engine with Compose V2; the CLI requires Rust 1.85 or later. |
| Paritok | Route coding-agent requests through a local proxy. | Self-hosted, Server | The project is public and licensed under Apache-2.0. The documented self-hosted quick start uses Python 3.11 or later and Ollama. |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 13 October 2026
- Original website title
- tastyeffectco/sandboxd · Official homepage title · retained in its source language