Releases

Release on Neural Inverse Cloud

We deploy incremental changes to Neural Inverse Cloud production multiple times per day.

How to release a change:

• Neural Inverse Cloud: the production branch deploys to all Neural Inverse Cloud regions. We frequently first release changes to Neural Inverse Cloud this way in order to monitor for potential failures before cutting an OSS release.
• OSS: (see below)

Release OSS

We cut new releases multiple times per week. We follow semantic versioning.

1. Make sure you are on main; run git status to make sure that you track origin:main.
2. Run pnpm run release, this:
   1. Bumps version numbers across packages
   2. Pushes commit to main
   3. Triggers release note on GitHub (please group the changes according to conventional commit type; see previous releases for reference)
3. GitHub Actions will automatically force push this commit to the production branch for a deployment to Neural Inverse Cloud (see above); no additional action necessary.

New features

When we release a new feature:

• Let customers know who asked for this or shared feedback; if you link Pylon threads to Linear issues, they are automatically changed to close the loop status when issue is marked as done in Linear (see product ops)
• Update the documentation
• Create a changelog post

---

---