About flag92
flag92 is a bilingual guide to open-source AI customer support stacks. We curate projects worth deploying, document combos that actually ship, write deploy guides from production scars, and publish benchmarks from real projects.
Why this site exists
Over the past two years open-source AI support (Chatwoot, Dify, FastGPT, RAGFlow, etc.) matured fast — but English material is dominated by vendor blogs, and Chinese material is scattered. Every time we helped a customer evaluate options, we re-researched 20+ posts to compare versions, licenses and gotchas. flag92 collects this work once so others don't have to.
We don't sell SaaS or closed-source plugins. The business model is simple: the "Hosted alternatives" page lists a handful of managed services (including one affiliate partner). If you click through and sign up we get a referral fee. Most of the site has nothing to do with them, and all alternatives are listed without ranking.
Content scale
Our principles
Only field-tested content
Every tool, solution and deploy guide comes from real-project experience. We don't write about what we haven't used.
Transparent about bias
Links, recommendations and comparisons disclose conflicts of interest. Sponsorships are labeled.
Updated regularly
Every content page carries updatedAt; stale entries are flagged or rewritten. Open-source evolves fast — outdated docs are worthless.
Welcoming pushback
Found an error or disagree? GitHub Issue or email — we respond publicly.
How content is sourced
- Every tool page is based on the project's docs plus our real deploy experience — we have at least docker-compose'd it
- Solution combos and industry playbooks come from real engagements (numbers anonymized, architecture generalized)
- Benchmark posts use the same environment, dataset and prompt across compared options
- Migration guides are notes from actual customer migrations
Contact
Need paid consulting, a tailored design, or training? Email us. We also run half-day stack-selection workshops for company teams.