A new tool called ShipNotes is making waves by using a small AI model to automatically generate clear, user-friendly release notes directly from your technical git logs. It runs efficiently on just one CPU core, saving developers time and effort.
Tired of writing release notes after every software update? We have exciting news for you! A brilliant new tool called ShipNotes is here to make your life much easier. It uses artificial intelligence to transform complex Git logs into clear, understandable release notes for everyday users, all while running on just a single CPU core.
Developers know this feeling well: you've merged the code, pushed the tag, and now it's time to tell users what's changed. The problem is, commit messages are often written for code reviewers who understand the technical details, not for the person using the final product. The result? Release notes that are either raw, hard-to-understand technical logs or simply skipped entirely due to time constraints. This leaves users confused and makes your update efforts less clear.
This is where ShipNotes intelligently steps in. You simply paste your commit messages, or point it at a public GitHub repository and a tag range, and ShipNotes does the rest. You'll get a Markdown changelog neatly grouped into «New», «Improved», and «Fixed» sections, with breaking changes flagged and Pull Request (PR) numbers linked. Plus, you receive a short 'What's new' paragraph ready for a tweet or to top an email, and a list of every commit it left out along with the reasons why. The tool supports output in either English or Simplified Chinese.
And the most exciting part? ShipNotes doesn't require massive computing power. It uses the Gemma 1B AI model, an open-weights model, running through llama.cpp on a single CPU core of a small Virtual Private Server (VPS). This means no API keys, no per-token bills, and if you choose to host it yourself, your private commit messages never leave your server.
Indie developers and open-source software maintainers can benefit from a hosted version of ShipNotes for about $3/month. This affordable cost is possible because the marginal cost of a release is virtually zero. It's a fantastic example of how AI can be made accessible and practical without the need for colossal resources.
You can try it yourself right now! Visit the official site and click «Write release notes» to run a sample, or switch to «Try it on a public repo» to test it on a public repository like simonw/llm. Please be a little patient; since it runs on one shared CPU core and processes one request at a time, a release with 15 commits might take 30 to 50 seconds. It's a great opportunity to experience the future of release notes.
Developers know this feeling well: you've merged the code, pushed the tag, and now it's time to tell users what's changed. The problem is, commit messages are often written for code reviewers who understand the technical details, not for the person using the final product. The result? Release notes that are either raw, hard-to-understand technical logs or simply skipped entirely due to time constraints. This leaves users confused and makes your update efforts less clear.
This is where ShipNotes intelligently steps in. You simply paste your commit messages, or point it at a public GitHub repository and a tag range, and ShipNotes does the rest. You'll get a Markdown changelog neatly grouped into «New», «Improved», and «Fixed» sections, with breaking changes flagged and Pull Request (PR) numbers linked. Plus, you receive a short 'What's new' paragraph ready for a tweet or to top an email, and a list of every commit it left out along with the reasons why. The tool supports output in either English or Simplified Chinese.
And the most exciting part? ShipNotes doesn't require massive computing power. It uses the Gemma 1B AI model, an open-weights model, running through llama.cpp on a single CPU core of a small Virtual Private Server (VPS). This means no API keys, no per-token bills, and if you choose to host it yourself, your private commit messages never leave your server.
Indie developers and open-source software maintainers can benefit from a hosted version of ShipNotes for about $3/month. This affordable cost is possible because the marginal cost of a release is virtually zero. It's a fantastic example of how AI can be made accessible and practical without the need for colossal resources.
You can try it yourself right now! Visit the official site and click «Write release notes» to run a sample, or switch to «Try it on a public repo» to test it on a public repository like simonw/llm. Please be a little patient; since it runs on one shared CPU core and processes one request at a time, a release with 15 commits might take 30 to 50 seconds. It's a great opportunity to experience the future of release notes.
