Matt Coles is an independent software publisher whose work is available through a public GitHub repository, offering a focused portfolio currently centered on developer tooling. The publisher's primary product, lgtmaybe, is a provider-agnostic pull request reviewer designed to work with hosted language models, locally run models, and any service exposing an OpenAI-compatible API. This flexibility means developers are not locked into a single vendor and can choose between cloud-based providers for convenience or local models for privacy, cost control, or offline use. The tool fits within the broader category of code review automation software, which assists engineering teams by analyzing proposed changes, flagging potential issues, and offering suggestions before code is merged. Typical use cases include streamlining review workflows in small teams, providing a first pass of automated feedback on pull requests, reducing the burden on senior reviewers, and maintaining consistent code quality standards across projects. Because it supports OpenAI-compatible endpoints, lgtmaybe can integrate with a wide ecosystem of inference providers and self-hosted model servers, making it suitable for organizations with strict data handling requirements as well as individual developers experimenting with different models. As a single-package publisher, Matt Coles exemplifies the open-source model of software distribution, where tools are shared publicly, iterated upon in the open, and made freely available to the development community. Users interested in this type of solution can explore the source code, configuration options, and documentation directly through the publisher's GitHub presence, where contributions and issue reporting are also typically welcomed.
Provider-agnostic pull request reviewer for hosted, local, and OpenAI-compatible models.
Details