Who runs this site
OSS Alternatives is written and operated by Yusuke Morinaga, an independent developer based in Japan (GitHub: @mori7ga2222, [email protected]). I build and self-host the kind of infrastructure described on this site as part of my day-to-day work on small SaaS and directory products, and I make the ranking and editorial calls on every page personally. This is a one-person publication, not a company front for a vendor, a paid review shop, or an SEO agency pretending to be a publisher.
The registered business mailing address is published in full on the Contact page, alongside the correction and takedown process.
Why this site exists
Most "alternatives to X" lists you find on the open web are either (a) marketing pages from a direct competitor, (b) automated scrapes with no opinion, or (c) outdated rankings from people who never actually deployed any of the tools. We wanted the in-between: opinionated tradeoff notes anchored to live GitHub data, kept fresh by a deterministic ETL pipeline, with the AI-assisted parts clearly labeled so you know what is editorial framing vs. what is upstream fact.
See the Methodology page for the editorial process this site applies in detail — where the data comes from, how projects are ranked, and what is reviewed by hand.
Where the self-hosting experience comes from
The cost ranges, migration timelines, and "stay on SaaS" caveats on this site are not pulled from vendor white papers. They are what the operator has repeatedly hit on their own infrastructure:
- Several years of running production workloads on $5–20/month VPS instances and small managed Postgres deployments
- Multiple SaaS-to-self-hosted migrations on adjacent properties — moving error tracking off a hosted vendor onto a self-run OSS stack, replacing a hosted ETL scheduler with cron + a Git-versioned pipeline, and replacing hosted site search with self-built Pagefind indexes (this site uses one). The analytics layer is intentionally a hybrid: privacy-friendly hosted (Cloudflare Web Analytics) for traffic counts plus self-hosted logs for everything else.
- Operating an automated content pipeline with nightly ETL, scheduled rebuilds, and an editorial polish loop powered by the Claude API
- The unglamorous parts: TLS auto-renewal that silently broke, a backup script that worked for 11 months and then didn't, a forgotten
postgresminor-version bump that pinned a restore drill, and the resulting incident notes that became the "operational cost" framing on Methodology
We surface this not to brag — it's a small operation — but because the most common feedback we get on OSS comparison content is "this looks like it was written by someone who has never actually run any of these tools." That criticism is fair for a lot of the genre. We try not to be that.
How we decide what counts as a serious alternative
Our objective inclusion thresholds are documented on the Methodology page, but the short version is: a SaaS comparison page is only indexed when it has at least 3 open-source alternatives, the top alternative has at least 800 GitHub stars, and the editorial intro is a real generated paragraph (not a templated stub). The why behind those specific numbers:
- ≥ 3 alternatives — fewer than 3 makes the comparison table less useful than just reading the top project's README. A one- or two-entry page is a redirect, not a directory entry.
- ≥ 800 stars on the top alternative — below this, the project is usually too early for production teams to evaluate seriously, and our pros/cons summary won't be reliable because we don't have enough community signal to lean on.
- ≥ 80-character intro — eliminates pages where the editorial pass failed and fell back to a one-line template. Those SaaS are not given a public comparison page until the intro clears this bar, so the index only contains pages that are actually worth reading.
AI usage policy
Per-SaaS intros, pros/cons framing, and FAQ phrasing are generated by Claude Sonnet 4.6 and Claude Haiku 4.5 via the Anthropic API or Claude Code. Repository facts (stars, license, language, last pushed) come straight from the GitHub public API and are not touched by any language model. The legal pages, this About page, the Methodology page, and per-alternative honest-tradeoff takes (when present) are written or line-edited by hand. The complete split — including what models can and cannot be trusted with — is on the Methodology page (AI vs. human roles).
Corrections and contact
Found a wrong license, an archived project we're still treating as active, or an alternative that should obviously be on a list and isn't? Email [email protected] with the page URL and what's wrong, or use the Contact page for the full intake format (correction, removal, partnership, bug). We acknowledge within 5 business days and aim to ship corrections within 14 days. Project maintainers requesting removal: see the verification steps on Contact.
Editorial independence and monetization
We are not affiliated with any vendor, project, or model listed on this site. The site is funded by a combination of disclosed affiliate links (primarily for VPS hosting and adjacent services that readers would likely buy anyway when self-hosting) and, where eligible, display advertising. Affiliate commissions never buy placement, ranking, or a more favorable summary — the full editorial-independence statement and the AI / human-roles split is on the Affiliate Disclosure, and the Privacy Policy covers how visitor data is handled.