For agent and assistant builders
Your agent can search the web. It cannot ask who to trust.
Search returns whatever ranks highest this week. It cannot tell you who a specific community has actually come to rely on, or which take earns respect from people who usually disagree. UpTrust keeps a graph of that, built from real votes between real people, and is opening it up so an agent can query it directly.
What an agent can ask
Three questions a popularity score cannot answer
Who is trusted on a topic
Ask for the people a given network actually relies on for a subject, ranked by earned trust rather than follower counts.
Who bridges two camps
Find the people and posts that earn respect from both sides of a disagreement, not just applause from one side.
Who someone should meet
Suggest an introduction based on real trust relationships, with the user's permission, instead of a semantic-similarity guess.
Already in the product
A person can already ask their own assistant
Early, and permissioned
This part is early, and we are opening it up step by step
An agent can only see what its user could see, and only after that user connects it. Nothing here is scraped. We are building this in the open rather than promising a finished platform before it is one.
“You have processes whereby power is transferred from human beings to AI systems, and you don't really have processes that go the other way…”
Build on a trust graph, not a scrape.
Sign up to see the graph your own trust network builds, and where an agent could take it from there.
Sign Up