A developer built a free desktop app that puts Claude, Codex, Gemini and Ollama in one window

What the developer built

As dev.to reported, a developer got tired of a coding day spread across terminal windows and built a tool to fix it. The result is SuperAi, a free, open-source desktop workspace that pulls several AI assistants into one place.

The tools it covers are Claude Code, ChatGPT Codex, Gemini, OpenCode, and local models run through Ollama. Most of these are command-line tools, or CLIs: programs you drive by typing text into a terminal rather than clicking in a window. Ollama is a program that runs AI models on your own computer instead of on a company's servers.

The author's complaint will be familiar to anyone who uses more than one assistant. They were switching tabs all day, copy-pasting context from one tool to another, and wishing they could hand the same prompt to two models at once to see which wrote cleaner code.

What you can do inside it

The post lists six main features. The first is the core idea: run and manage the different assistants side by side, without a pile of terminal tabs.

• Parallel comparison: send one prompt to several agents at once and compare their reasoning and code as the answers arrive.
• Multi-agent orchestration: chain agents into automated workflows. The author's example is one model drawing up the plan while another writes and reviews the code.
• Terminal and web preview: run commands inside the app and preview web output without leaving it.
• MCP support: connect servers that use the Model Context Protocol, an open standard for linking AI apps to outside tools.
• Voice input: speak your prompts, in English or Thai.

These are the developer's own descriptions. The post does not offer benchmarks or independent testing, so how well each feature works is something you would have to judge by trying it.

Why compare models side by side

Sending one prompt to several models is a practical habit more than a gimmick. Different assistants often handle the same request differently, and the only way to know which suits your code is to see the outputs together. Doing that by hand means pasting the same text into several windows and lining up the results yourself.

The chaining idea goes a step further. Instead of choosing one assistant, you give each a job: one plans, another writes, another reviews. The post presents this as automated workflows, though it does not say how much setup they need or how reliable they are.

Including Ollama matters as well. A model on your own machine can sit beside the hosted ones, so you can compare a local option with a paid service in the same view.

How to get it

The developer says SuperAi is free and open-source. It is on the Microsoft Store, which the post says is meant to make installation safe on Windows, with no SmartScreen warnings. SmartScreen is the Windows feature that warns you before running software it does not recognize. The code is also on GitHub.

If you prefer the command line, the post gives a Winget install: winget install 9NZH79QL515K. Winget is the package installer built into Windows Terminal. The post talks only about Windows, so do not assume versions exist for other systems.

What to check before trusting it

The post leaves open how SuperAi stores credentials or what it sends over the network. That matters. An app that wraps several AI tools sits right next to your API keys, which are the secret codes that let software use your accounts, and next to your source code.

Open source helps here, because anyone can read the code. A reasonable first step is to look through the repository for how keys are handled and what network calls the app makes. Then try it on a project that holds nothing sensitive.

It is also an early, one-developer project. Whether it keeps pace as the underlying tools change is an open question. Its value will depend on how well it handles each assistant's updates and how actively the author maintains it.