Skip to main content

Applied AI, built for real work

OneTest is an open-source product studio building tools for people and AI agents. Our projects live where software work happens—in repositories, GitHub, VS Code, native desktop applications, and the browser. We build focused tools that keep plans, artifacts, and decisions visible to the people doing the work.
A bento layout showing OneTest tools for testing, desktop AI, presentations, and project executionOneTest octopus logo

OneTest connects repository-native quality, desktop AI, presentations, and project execution.

Explore our projects

Follow the work, try a project, open an issue, or contribute on GitHub.

OneTest TMS

Start with GitHub-native test management for humans and agents.

Projects

OneTest TMS

GitHub-native test management. Keep test assets in files, manage runs through GitHub, and let agents operate the workflow through MCP.

DeepSeek Harness Desktop

A native macOS shell for DeepSeek Harness with managed installs, plugins, MCP servers, notifications, an editor, and a real browser.

Octodeck

Deck-as-code for people and agents. Build typed, interactive presentations in TypeScript and export them as self-contained HTML.

Octobots

A repository-native project board for AI coding teams, built from durable YAML plans, acceptance criteria, and completion gates.

How we build

Repository-native

Important work stays in files and standard developer systems instead of disappearing into a separate database.

Human and agent operated

People and AI agents work from the same plans, artifacts, acceptance criteria, and history.

Focused and composable

Each project solves a concrete problem and connects through Git, MCP, plugins, and familiar development tools.

Open and inspectable

Our tools, formats, decisions, and implementation are public, portable, and open to contribution.

From software quality to Applied AI

OneTest began with AI-assisted software quality. That work continues in OneTest TMS, now being rebuilt around GitHub-native workflows. Along the way, we found the same problem across software work: agents need durable context, explicit acceptance criteria, inspectable artifacts, and tools that fit the way teams already work. OneTest now explores that problem across planning, development, presentations, and desktop AI.

Built in public

These projects are active experiments as well as useful tools. We share the work early, document the tradeoffs, and improve it in the open.

View the GitHub organization

Browse source code, releases, issues, and project history.

Read how we work

Learn the principles connecting the OneTest projects.