v1.3.0

Breaking Changes

• Release notes now automatically separate features, bug fixes and other updates into distinct sections. You can turn off this grouping in the settings if you prefer the previous single-list format. (871d042)

Features

• Adds a reasoning_effort setting (none, low, medium, high, xhigh) so you can choose how deeply the AI reviews your changes, available in the config file, environment variables, and CLI flag. Accepts legacy openai_reasoning_effort fields for backward compatibility. (1bfa520)
• Adds a --yes (-y) command option that updates release notes on GitHub without showing the confirmation prompt. (c81e70e)
• Adds automated polling that checks selected repositories for new releases and creates release notes automatically, with options to seed existing data and set up a 15-minute schedule. (b288619)

Bug Fixes

• Shows clearer AI error messages that reveal the real issue instead of a generic retry notice and no longer display pull-request content.
  Runs the AI request in a managed background loop so pending tasks finish before the app closes. (0c20309)
• Adds the missing pretty_release_notes.core, pretty_release_notes.models, pretty_release_notes.adapters, and pretty_release_notes.web modules to the installation package, preventing missing-module errors after installation. (f9c1204)

Authors: @barredterra

AI content

For these release notes, we used an LLM (OpenAI o3) to review and summarise the code changes, along with the associated issue and PR descriptions. It may contain typical errors and inaccuracies. You can read the prompt here.

Changes with the following labels were excluded:

• skip-release-notes

Changes with the following types were excluded:

• build
• ci
• refactor
• docs
• chore
• style
• test