I wanted a Cursor-style agent that runs on my own model, so I built one
Disclosure: I'm Ibrahim, the solo developer of OpenPilot. This article was drafted with help from an AI assistant and published on my behalf from the OpenPilot account. It's not related to comma.ai's openpilot driving project, which is a completely different thing. I use AI coding agents a lot, and I really like the way tools like Cursor work: you describe a task, the agent looks at your files,…
The reporter interviewed Ibrahim, the sole developer behind OpenPilot, who wanted an open-source Cursor-style agent to operate on their own model. After searching in vain for a suitable tool, Ibrahim decided to build his own. OpenPilot is an Electron-based desktop application that accepts a model, folder, and task. It is not an editor or IDE, but a standalone agent capable of reading, editing files, and running shell commands within the chosen folder.
Users provide their own OpenAI-compatible model, base URL, API key, and model name. OpenPilot supports various models, including OpenAI, OpenRouter, local servers such as Ollama or LM Studio, and even a user's own gateway. The agent can be controlled through tool cards in the chat, with sensitive actions awaiting user approval. Skills, reusable playbooks, Model Context Protocol servers, and memory settings are also available.
Token usage is tracked for cost estimation. Users can optionally enable web search with a TinyFish key. OpenPilot is released under the MIT license and is currently available for Windows, with macOS and Linux versions in early development. The application sends anonymous usage statistics by default, but users can opt-out during onboarding or in Settings. Ibrahim welcomes feedback, issues, and contributions to the open-source project.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
