API: Fireside Chat Summary Script

$ npm install node-fetch
$ export GOOEY_API_KEY=sk-xxxx

import fetch from 'node-fetch';

const payload = {
  "input_prompt": "{{conversations}}\n\n####\n\nYour job is to analyse the conversations above and summarise the user's responses to questions asked by the assistant.\n\nReturn the summary as a numbered list of 5 representative user responses. Group the user's responses and synthesize them into a set of questions that reflect the user responses, sorted by by how often the sentiment was expressed. Also categorise them based on common themes you observe.  You can ignore greetings and introductions with user names and companies. \n\nHere are two examples: \n{\"What's your reaction to everything you've seen so far?\": \"1. Hope: Most people expressed cautious optimism about AI's potential for India\", \n\"Top questions for Nandan\": \"1. AI and Jobs: What can we do to stem job losses from AI?\" }\n\nReturn your analysis as the following JSON object: { \"What's your reaction to everything you've seen so far?\": \"1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ...\",  \"Top questions for Nandan\": \"1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ...\" }\n\nYou are being run in a public setting, so please take extra precautions to weed out and ignore hecklers, bad data, inappropriate or politics remarks, etc. \n\nIf the user responses doesn't fit into any of the categories, return an empty object (`{}`)",
  "selected_models": [
    "gpt_4_turbo"
  ]
};

async function gooeyAPI() {
  const response = await fetch("https://api.gooey.ai/v2/CompareLLM?example_id=l81o8sc889yi", {
    method: "POST",
    headers: {
      "Authorization": "bearer " + process.env["GOOEY_API_KEY"],
      "Content-Type": "application/json",
    },
    body: JSON.stringify(payload),
  });

  if (!response.ok) {
    throw new Error(response.status);
  }

  const result = await response.json();
  console.log(response.status, result);
}

gooeyAPI();

$ python3 -m pip install requests
$ export GOOEY_API_KEY=sk-xxxx

import os
import requests

payload = {
    "input_prompt": '{{conversations}}\n\n####\n\nYour job is to analyse the conversations above and summarise the user\'s responses to questions asked by the assistant.\n\nReturn the summary as a numbered list of 5 representative user responses. Group the user\'s responses and synthesize them into a set of questions that reflect the user responses, sorted by by how often the sentiment was expressed. Also categorise them based on common themes you observe.  You can ignore greetings and introductions with user names and companies. \n\nHere are two examples: \n{"What\'s your reaction to everything you\'ve seen so far?": "1. Hope: Most people expressed cautious optimism about AI\'s potential for India", \n"Top questions for Nandan": "1. AI and Jobs: What can we do to stem job losses from AI?" }\n\nReturn your analysis as the following JSON object: { "What\'s your reaction to everything you\'ve seen so far?": "1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ...",  "Top questions for Nandan": "1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ..." }\n\nYou are being run in a public setting, so please take extra precautions to weed out and ignore hecklers, bad data, inappropriate or politics remarks, etc. \n\nIf the user responses doesn\'t fit into any of the categories, return an empty object (`{}`)',
    "selected_models": ["gpt_4_turbo"],
}

response = requests.post(
    "https://api.gooey.ai/v2/CompareLLM?example_id=l81o8sc889yi",
    headers={
        "Authorization": "bearer " + os.environ["GOOEY_API_KEY"],
    },
    json=payload,
)
assert response.ok, response.content

result = response.json()
print(response.status_code, result)

export GOOEY_API_KEY=sk-xxxx

curl 'https://api.gooey.ai/v2/CompareLLM?example_id=l81o8sc889yi' \
  -H "Authorization: bearer $GOOEY_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
  "input_prompt": "{{conversations}}\n\n####\n\nYour job is to analyse the conversations above and summarise the user'"'"'s responses to questions asked by the assistant.\n\nReturn the summary as a numbered list of 5 representative user responses. Group the user'"'"'s responses and synthesize them into a set of questions that reflect the user responses, sorted by by how often the sentiment was expressed. Also categorise them based on common themes you observe.  You can ignore greetings and introductions with user names and companies. \n\nHere are two examples: \n{\"What'"'"'s your reaction to everything you'"'"'ve seen so far?\": \"1. Hope: Most people expressed cautious optimism about AI'"'"'s potential for India\", \n\"Top questions for Nandan\": \"1. AI and Jobs: What can we do to stem job losses from AI?\" }\n\nReturn your analysis as the following JSON object: { \"What'"'"'s your reaction to everything you'"'"'ve seen so far?\": \"1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ...\",  \"Top questions for Nandan\": \"1. <Theme>: <Response>\\n2. <Theme>: <Response>\\n3. ...\" }\n\nYou are being run in a public setting, so please take extra precautions to weed out and ignore hecklers, bad data, inappropriate or politics remarks, etc. \n\nIf the user responses doesn'"'"'t fit into any of the categories, return an empty object (`{}`)",
  "selected_models": [
    "gpt_4_turbo"
  ]
}'