Ask a question out loud or type it — Rasepi pulls the answer straight from your documentation, with links to the exact pages it came from. Voice or keyboard, your choice. And because Rasepi tracks freshness, the AI only uses content it can actually trust.
Most AI search tools will happily return a confident answer based on a document that hasn't been updated in two years. You get an answer. You just can't tell if it's still right.
A conversational interface — voice or text — on top of your knowledge base, with freshness as a first-class signal.
Say "What's our policy on remote work?" out loud or type "How do I set up the staging environment?" Either way, no special syntax needed. Just a natural question.
The AI searches across all docs in your workspace (respecting permissions). It filters by freshness score, so stale or expired content doesn't pollute the results.
The response includes a clear answer plus links to every source document. Each source shows its freshness score, so you know at a glance how trustworthy the information is.
If the best available answer comes from a doc with a low trust score, you'll see a warning. The AI is honest about it: "This is the best match I found, but the source was last reviewed 6 months ago."
The problem with most RAG (retrieval-augmented generation) setups is that they treat all documents equally. A policy from 2021 and a runbook updated yesterday have the same weight.
Rasepi changes that. Because every document already has a trust score, the AI has something most systems don't: a built-in signal for how much to trust each source. A doc scoring 95 is given more weight than one scoring 30. Simple, but it changes everything.
This is the advantage of building freshness into the platform from day one. It's not bolted on. It's the foundation that makes every other feature more reliable.
Ask questions that span multiple hubs. The AI searches across your entire workspace, not just one silo.
The AI only surfaces content you have access to. Private hub docs won't leak into answers for people who shouldn't see them.
Like translation, the AI backend will be pluggable. Use OpenAI, Anthropic, or a self-hosted model. Your choice, your API key.
Ask in German, get an answer sourced from English docs (and vice versa). The AI can search across all language versions of your content.
Ask your docs anything — speak it or type it. Freshness scoring ensures the AI only uses content that's been reviewed and is still current.
Rasepi is in private beta. We're inviting teams in waves.