other
Agora takes a claim or prediction and returns a verdict of true, false, or uncertain, backed by a confidence score rather than a flat answer with nothing behind it. Behind that verdict is an argument graph built from credibility-scored evidence, where opposing AI debaters argue their positions with cited reasoning rather than a single model asserting an answer on its own and expecting it to be trusted without support. Neotask can submit a claim to Agora and bring back not just the verdict but the reasoning structure behind it, so you see how the confidence score was reached instead of taking it on faith from a single line of output with no trail behind it. This suits anyone who needs to check a claim before repeating it: a fact someone shared in a meeting, a prediction floating around online, a statement worth verifying before it goes into something public with your name attached to it. Because the evidence is credibility-scored and the debate is cited, the output reads closer to a transparent argument you can walk through yourself than a single rating pulled from an opaque process with no way to check the reasoning behind it. That structure also means two people looking at the same verdict can trace the same argument back to its sources instead of arguing about whether the answer itself can be trusted.
| submit_claim | Sends a claim or prediction to Agora for evaluation. This is the entry point for every other lookup here. |
| get_verdict | Returns the true, false, or uncertain result along with its confidence score. Gives a quick answer without the full argument graph. |
| get_argument_graph | Retrieves the underlying structure of evidence and debate behind a verdict. Useful when the verdict alone isn't enough to act on. |
| get_cited_evidence | Returns the credibility-scored sources cited on each side of the argument. Lets you check the sources directly rather than trusting the summary alone. |
An editor has a statistic in a draft newsletter she isn't fully sure is accurate, and the issue goes out in an hour. She asks Neotask to submit the claim to Agora, and the agent returns a verdict along with the confidence score and the evidence behind it. She either keeps the line, rewrites it, or cuts it based on what came back, with enough time to spare before the send. It's a small step that saves her from a correction email the next day, and one she now runs on any statistic she can't personally verify.
Two colleagues disagree over whether a prediction about a market trend holds up, and neither one is budging in the conversation. One asks Neotask to submit the prediction to Agora rather than argue it out further over email. The agent returns the verdict and the argument graph, and both sides look at the same cited evidence instead of repeating their own positions back and forth. The disagreement ends with a shared reference point instead of a stalemate that would have dragged on for another week, and the two move on to planning around the actual answer.
Agora returns true, false, or uncertain for a submitted claim or prediction, along with a confidence score attached to that verdict.
Through an argument graph built from credibility-scored evidence, where opposing AI debaters argue their positions with cited reasoning.
You can ask for the argument graph itself, not just the verdict, to see the evidence and debate behind it.
Any claim or prediction you want checked, from a shared fact to a forward-looking prediction.
Yes, the evidence in the argument graph is credibility-scored rather than treated as equally reliable.
It can at least give both sides the same cited evidence and reasoning to look at, rather than each person arguing from their own memory of the facts.
Yes, the argument graph behind the score is available on request, so the confidence figure doesn't stand alone without context.
Yes, resubmitting is reasonable if new evidence has emerged since the last check, since the debate draws on the evidence available at the time.
Yes, predictions are explicitly part of what can be submitted, alongside claims about something that has already happened.