Login Book a demo

Monitoring

List verdicts

See how one conversation did against each rule, and why.

POSThttps://api.ollive.ai/api/verdicts/query/v1
Safe to retryv1

How it works

A verdict is the result of checking one rule in one conversation: pass, fail, or indeterminate when there wasn’t enough information to decide. Every rule that was checked has a verdict, including the ones that passed.

Each verdict lists the facts Ollive found in the conversation, with quotes. A failed verdict also has a proof: the reasoning, and the text of the rule’s source.

Authorization

  • Authorizationheaderrequired

    Your API key, after the word Bearer, like this: Bearer sk-…. Create a key.

Request body

application/json

  • session_refstringrequired

    Your ID for the conversation.

  • verdictstring

    Only verdicts with this result.

    passfailindeterminate
  • directive_idstring

    Only this rule.

  • binding_text_idstring

    Only rules from this source.

  • deployment_idstring

    Only this deployment.

  • sourcestring

    Only needed if you used the same session_ref for a conversation Ollive collected and one you sent.

    PLATFORMAPI

Response

The verdicts. An empty list if there are none.

  • dataobject[]required

    One verdict per rule checked.

    • idstring (uuid)required

      The verdict’s ID.

    • session_refstringrequired

      Your ID for the conversation.

    • sourcestringrequired

      PLATFORM if Ollive collected it, API if you sent it.

      PLATFORMAPI
    • versionintegerrequired

      Which review this is. Usually 1.

    • deployment_idstring (uuid)required

      The deployment.

    • directive_idstringrequired

      The rule.

    • binding_text_idstringrequired

      The law, regulation or contract the rule comes from.

    • pack_idstringrequired

      The rule pack it belongs to.

    • verdictstringrequired

      The result.

      passfailindeterminate
    • appliedbooleanrequired

      Whether the rule applied to this conversation.

    • severitystringrequired

      How serious the rule is.

      criticalhighmediumlow
    • citationsstring[]required

      References to the source, such as a section number.

    • factsobject[]required

      What Ollive found in the conversation to reach the verdict.

      • fieldstringrequired

        The fact’s name.

      • questionstringrequired

        The question Ollive answered.

      • answerstringrequired

        The answer.

        truefalseunknown
      • rationalestringrequired

        Why.

      • evidenceobject[]required

        Quotes from the conversation that support the answer.

        • turn_indexintegerrequired

          The turn the quote is from.

        • speakerstringrequired

          Who said it.

          agentcallertool
        • start_quotestringrequired

          Where the quote starts.

        • end_quotestringrequired

          Where it ends.

    • proofobject

      Why the verdict failed. Only on failed verdicts.

      • rationalestringrequired

        The reasoning, in plain words.

      • decisive_fieldsstring[]required

        The facts that decided it.

      • clausesobject[]required

        The text of the rule’s source.

        • citationstringrequired

          The reference, such as a section number.

        • textstringrequired

          The source text, word for word.

        • source_urlstringrequired

          Where to read it.

    • session_started_atstring (date-time)

      When the conversation started.

    • session_ended_atstring (date-time)

      When it ended.

    • verified_atstring (date-time)

      When the check finished.

← List violationsGet metrics →