Havoptic Methodology

This page documents how Havoptic collects, processes, and publishes data about AI coding tool releases. The site is maintained by Scott Havird.

Tracked tools and upstream sources

Fetch cadence

Release fetchers run once per day via scheduled GitHub Actions workflows. When a new release is detected, it is appended to releases.json, a new infographic is auto-generated, and a notification is dispatched to subscribers and push subscribers. The full pipeline typically completes within 30 minutes of a detected upstream change.

AI-assisted analysis and human review

Release summaries and blog posts (weekly digests, monthly comparisons, tool deep-dives) are drafted using large language models, then reviewed by Scott Havird before publication. Infographics are auto-generated from the release metadata. Raw release data and version strings are always taken verbatim from the upstream source — no LLM-generated content is introduced to the structured data feeds.

Corrections policy

Factual errors should be reported to [email protected] or via a GitHub issue at scotthavird/havoptic.com. Confirmed corrections are applied to the next daily regeneration cycle; the dateModified on affected pages updates accordingly.

Data license

Release data and the feature matrix are published under Creative Commons Attribution 4.0. You may reuse the data for any purpose with attribution to Havoptic.

Conflict of interest

Havoptic has no financial relationship with any tracked tool vendor. The site earns no affiliate or referral revenue. Coverage order is alphabetical; release priority is first-seen.

About Havoptic · API documentation