Basic agent - Trodo

The smallest useful agent: wrap one LLM call, capture input and output, attach feedback after the user reacts.

Node.js
Python

import trodo from 'trodo-node';
import OpenAI from 'openai';

trodo.init({ siteId: process.env.TRODO_SITE_ID });
const openai = new OpenAI();

export async function answer(userId, conversationId, question) {
  const { result, runId } = await trodo.wrapAgent(
    'support-bot',
    async (run) => {
      run.setInput({ question });

      const r = await openai.chat.completions.create({
        model: 'gpt-4o-mini',
        messages: [
          { role: 'system', content: 'You are a concise support assistant.' },
          { role: 'user', content: question },
        ],
      });
      const text = r.choices[0].message.content;

      run.setOutput({ answer: text });
      return text;
    },
    { distinctId: userId, conversationId },
  );

  // When the user clicks thumbs-up on your UI:
  await trodo.feedback(runId, { satisfaction: 'positive', rating: 5 });
  return result;
}

import trodo, os
from openai import OpenAI

trodo.init(site_id=os.environ["TRODO_SITE_ID"])
client = OpenAI()

def answer(user_id, conversation_id, question):
    with trodo.wrap_agent("support-bot", distinct_id=user_id, conversation_id=conversation_id) as run:
        run.set_input({"question": question})

        r = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
                {"role": "system", "content": "You are a concise support assistant."},
                {"role": "user", "content": question},
            ],
        )
        text = r.choices[0].message.content
        run.set_output({"answer": text})

        trodo.feedback(run.run_id, satisfaction="positive", rating=5)
        return text

What gets captured

Run — support-bot, status ok, duration end-to-end, tokens + cost rolled up from the LLM span.
Span — one kind='llm' span with model, provider, input/output tokens, cost, and temperature auto-filled by the OpenAI auto-instrumenter.
Feedback — one positive / rating 5 row linked to the run.

​What gets captured

​See also

What gets captured

See also