Cited intelligence for DXP teams
Every DXP answer, cited at the source.
Sitecore, Optimizely, Umbraco, Contentful, Contentstack. Every claim traces to a real vendor doc at a specific version. No invented answers, no blended versions.
617 records · 9 version tracks · no source, no record
No source, no record.
Every answer, verifiable.
The problem
AI gives DXP teams confident soup.
Migration, upgrade, and implementation teams re-solve the same operational questions on every engagement. Vendor docs differ per version and per platform. The AI tools they are offered blend it all into fluent, wrong answers.
Twofold answers once, cited, and pinned to a version. The knowledge base is the source of truth; the product never invents one.
Why it is different
Answers you can verify
Every claim cites a real vendor doc at a specific version. Follow the link and see it yourself.
Never blended
umbraco/13-lts is umbraco/13-lts, not "Umbraco". The version track is always stated, never mixed.
Honest gaps
When the knowledge base has no record, it says so instead of inventing one. Missing coverage is reported as missing.
Built by operators
Thirty years inside Sitecore, Optimizely, and Umbraco. We know which upgrade steps carry hidden effort.
How it works
Ask in plain language
Search or ask about a migration, upgrade, or implementation, the way you would ask a senior engineer.
Get cited answers
Every response carries its record id, exact version track, confidence, and the vendor source URLs.
Verify it yourself
Follow the citations to the source. No black box, no blend of versions, no made up specifics.
The provable layer
One small stack, built so you can check the work.
Questions
Before you ask.
What is dxpwise? +
Cited intelligence for DXP teams. A CLI and web workspace that search and chat over the dxp-kb knowledge base, with every claim traced to a vendor doc at a specific version.
Which platforms are covered? +
Sitecore XM Cloud and XP, Optimizely CMS, Umbraco, Contentful, and Contentstack, across nine version tracks, including the upgrade paths teams hit most.
How do answers stay correct? +
Three invariants, enforced mechanically. No source, no record. Never blend versions. Honest gaps. Every change lands through review against hand-written ground truth.
What happens when there is no answer? +
It says so. The system reports missing coverage instead of generating a fluent guess. That is the point.
Do I need my own model key? +
Search runs with no model at all. The cited agent can use OpenRouter, OpenAI, or Anthropic, either ours or your own key.
What is the migration and upgrade assessment? +
A scoped deliverable built on the knowledge base. It turns your migration or upgrade plan into a cited, effort-scored assessment instead of a re-research project.
Do we own what we get? +
Yes. The knowledge you build and the assessments you receive are yours. The platform is source available and runs in your environment.
Is this a replacement for our engineers? +
No. It removes the re-research so your engineers spend their time on the parts that need judgment.
Request access
Cited answers for your next migration.
Tell us what you are migrating or upgrading. We will show you the cited answer, and if the knowledge base has no record, we will say so.