Hi,

Great job on the new website, I really like it!

+1

François
[email protected]
[email protected]

Le 09/02/2026 à 20:49, Serge Huber a écrit :
Hi all,

As you might have noticed, I've been working on quite a few things for
Unomi 3 lately, and one thing that's been bugging me for a while is
the state of our website. I think now that we have a 3.0 release out
the door, it's a good time to bring the site up to the same level.

I put together (yes AI was involved here too, sorry :)) a pretty
comprehensive website redesign, and I'd really like to get your
feedback before merging anything.

JIRA: https://issues.apache.org/jira/browse/UNOMI-932
Branch: UNOMI-932-website-modernization
PR: https://github.com/apache/unomi-site/pull/40
Staging preview: https://unomi-v3-site.netlify.app/


Why this is needed
==================

The biggest complaint I hear from users is that they can't find things
on the website — documentation, mailing lists, how to get started.
People just give up. I think we can do a lot better.

Beyond that, the world has changed quite a bit since the site was last
updated. CDPs are no longer just about web personalization — they can
also be the data backbone for AI agents, LLMs, recommendation engines
and real-time decisioning. Our website doesn't communicate any of
this, which means we're basically invisible to a whole generation of
developers who are looking for exactly what Unomi offers.

There are also a bunch of smaller issues that accumulated over time:
- The tutorial page still references Unomi 1.x/2.x patterns
- The testing guide has placeholder TODOs and content that was
actually copy-pasted from Apache Beam (references to Dataflow, Flink,
Spark, dead Jenkins links…)
- Team affiliations are outdated
- Several contribute pages have encoding issues (mojibake characters)
- No mention of the BOM we introduced in Unomi 3
- No SEO, no structured data, no sitemap, no AI-friendly metadata
- Video pages didn't look very appealing (ASF web guidelines)
- Company listings were in a fixed order which could look like we're
favoring certain vendors

What I changed
==============

The full diff is about 53 files, ~6,000 lines added and ~3,300
removed. Here's a summary:

Infrastructure:
- Bootstrap 4 → 5.3.8, Jekyll 4.2.0 → 4.4.1
- Mermaid.js v11 for interactive diagrams (MIT licensed, ASL v2 compatible)
- Complete CSS rewrite with centralized design tokens — 40+ CSS custom
properties, zero hardcoded colors in the HTML
- Docker build workflow using bretfisher/jekyll images (the old
jekyll/jekyll image is deprecated), outputs to target/site for Maven
publish compatibility
- Copyright updated to 2026

Homepage:
- Animated hero SVG showing the Unomi data flow (web, mobile, CRM, IoT
→ Unomi → segments, AI agents)
- New "CDP in the AI Era" section positioning Unomi as the data
foundation for AI agents and personalization engines
- Architecture Mermaid diagram that includes AI agent outputs
- Ecosystem marquee with randomized company order on each page load
(ASF neutrality)
- "What's New" timeline combining releases, articles and events —
capped at the 10 most recent

New and redesigned pages:
- New Integrations page — Products & Platforms, Extensions, Commercial
Support with randomized card order; includes Coming Soon cards for
Inoyu SaaS and Unomi UI (open source ASLv2)
- Use Cases — reordered with AI & ML first, 4 use cases each with a
Mermaid diagram
- Tutorial — completely rewritten for Unomi 3.0
- Resources — video thumbnails now hosted locally (no YouTube
tracking), article summaries added

Community and contribute:
- Community index — redesigned to be more appealing
- Team — redesigned from table to responsive card grid; updated
affiliations (Serge Huber → Inoyu, Jean-Baptiste Onofré → Dremio,
several others marked as unknown — please correct me if any are
wrong!)
- Events — rebuilt with all 12 talks/videos organized chronologically
with local thumbnails
- All 11 contribute sub-pages modernized (layout and CSS only — text
content was strictly preserved)
- Testing guide — 3 TODOs filled in with accurate content based on the
actual source code (Pax Exam, Karaf, Failsafe, GitHub Actions CI);
removed the Apache Beam boilerplate that had never been adapted
- Dependencies guide — added a new section documenting the BOM
(unomi-bom and unomi-bom-artifacts) with usage examples
- Fixed mojibake encoding issues on several pages

SEO & AI engine optimization:
- JSON-LD structured data (Schema.org) on all pages
- Open Graph and Twitter Card meta tags
- robots.txt and sitemap.xml
- llms.txt and llms-full.txt — plain-text project summaries for LLM crawlers

What did NOT change
===================

- No changes to the Unomi software itself or the hosted documentation/manual
- Contribute page text content was preserved exactly as approved —
only layout and CSS were modernized
- All existing URLs are maintained

Feedback requested
==================

Please take a look at the staging preview and let me know what you think :)

1. Content accuracy — Are team affiliations correct? Any descriptions
that misrepresent Unomi?
2. ASF policy compliance — Anything that might conflict with ASF guidelines?
3. Missing content — Pages or topics we should add?
4. Look and feel — General impressions, mobile, readability?

I'm happy to iterate based on your feedback.

Best regards,
   Serge Huber.
   Apache Unomi PMC Chair

Reply via email to