policybazaar.ai assistant

policybazaar.ai team · Chat · In-house

Live

Conversations

521

from files

Scoring new traffic

Every 15 min

last ran just now

Uploaded · 521 conversations from files

Upload another file

Up to 5,000 conversations per file. Not masked on the way in (the SDK masks in the app); Dvar masks names and numbers before any model reads them.

One conversation per line: {"id": "c1", "messages": [{"role": "user", "content": "…"}, {"role": "assistant", "content": "…"}]}

Go live

Connect the agent and every new conversation is scored automatically, every 15 minutes. No more exports.

No key yet. One key per workflow, for the gateway and the SDK.
  1. 1

    Point the agent's OpenAI-compatible client at Dvar

    from openai import OpenAI
    client = OpenAI(base_url="http://localhost:4000/v1", api_key="<your Dvar key>")
  2. 2

    Send each conversation's id, so its turns group into one conversation

    client.chat.completions.create(model="dvar-fast", messages=messages,
        extra_body={"metadata": {"session_id": conversation_id}})
  3. 3

    Conversations appear here within about 15 minutes.

Compare limits