LLM Insights API

Ingest and analyze AI-driven insights directly. By integrating with this API, you can capture LLM responses, user interactions, and metadata to monitor your application in real-time

The OpenAPI integration provides a direct way to ingest and manage insight data, enabling seamless integration for developers looking to analyze interactions and capture insights directly. This API supports the ingestion of OpenAI ChatCompletion responses and other metadata, empowering you with real-time analytics and insights.

Send logs to Requesty

Send chat messages, responses, and meta data to Requesty.

PUThttps://ingestion.requesty.ai/insight

Authorization

bearerAuth

Body

application/json

response*object

OpenAI ChatCompletion response

messages*one of

The actual messages passed to the OpenAI client

stringarray of stringarray of object

templateone of

Optional prompt template pre-formatting

stringarray of stringarray of object

inputsobject

Optional input variables for the template

argsobject

Optional model arguments for tracking

user_idnullable string

Optional user identifier

metaobject

Optional metadata for tracking

Response

200400401404500

Success

Body

application/json

successboolean

Example: true

Request

JavaScriptcURL

const REQUESTY_API_KEY = 'your-requesty-api-key';
const REQUESTY_API_URL = 'https://ingestion.requesty.ai/insight';

const sendToRequesty = async (chatResponse, messages, meta) => {
  await fetch(REQUESTY_API_URL, {
    method: 'PUT',
    headers: {
      Authorization: `Bearer ${REQUESTY_API_KEY}`,
      'Content-Type': 'application/json',
    },
    body: JSON.stringify({
      response: response,
      messages: messages,
      meta: meta,
    }),
  });
};

Response

200

{
  "success": true
}

Getting started

1. Register on the Platform

• Visit the platform and register for an account.

• After registration, create an organization in the dashboard.

2. Create an API Key

• Navigate to the Insights API page in your account.

• Generate an API key. This API key will be used to authenticate your requests.

• Note: The same API key can be used for both Client Insights and Voice Insights. The only difference is the ingestion endpoint.

3. Set up the API integration

Base URL

All API requests should be made to the following base URL:

https://ingestion.requesty.ai/

Authentication

The AInsight API requires Bearer Token Authentication for secure access. Include your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

---

Examples

Send to Requesty example

const response = { choices: [{ message: { content: 'Hello world!' } }] };
const messages = [{ role: 'user', content: 'Say hello!' }];
const meta = { timestamp: new Date().toISOString() };

sendToRequesty(chatResponse, messages, meta).then(() => {
  console.log('Data successfully sent to Requesty Insights!');
});

cURL example

curl -X PUT https://ingestion.requesty.ai/insight \
  -H "Authorization: Bearer your-requesty-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "response": {
      "choices": [
        {
          "message": {
            "content": "This is a test response"
          }
        }
      ]
    },
    "messages": [
      { "role": "user", "content": "Say this is a test" }
    ],
    "args": { "model": "gpt-4o-mini" },
    "meta": {
      "user_id": "user-12345"
    }
  }'

OpenAI-Requesty combination example

import { OpenAI } from 'openai';

const openai = new OpenAI({
  apiKey: 'your-openai-api-key',
});

const REQUESTY_API_URL = 'https://ingestion.requesty.ai/insight';
const REQUESTY_API_KEY = 'your-requesty-api-key';

interface InsightPayload {
  response: object;
  messages: { role: string; content: string }[];
  args: { model: string };
  meta: { user_id: string };
}

async function sendInsight(payload: InsightPayload): Promise<void> {
  const response = await fetch(REQUESTY_API_URL, {
    method: 'PUT',
    headers: {
      Authorization: `Bearer ${REQUESTY_API_KEY}`,
      'Content-Type': 'application/json',
    },
    body: JSON.stringify(payload),
  });

  if (!response.ok) {
    throw new Error(`Error capturing insight: ${response.statusText}`);
  }

  console.log('Insight captured:', await response.json());
}

async function main() {
  const messages = [{ role: 'user', content: 'Say this is a test' }];

  const chatResponse = await openai.chat.completions.create({
    model: 'gpt-4o-mini',
    messages: messages,
  });

  const payload: InsightPayload = {
    response: chatResponse,
    messages: messages,
    args: { model: 'gpt-4o-mini' },
    meta: { user_id: 'user-12345' },
  };

  try {
    await sendInsight(payload);
  } catch (error) {
    console.error(error);
  }
}

main();

---

Use Cases

• Real-Time Insights: Capture and analyze AI interactions in real-time.

• Auditing and Compliance: Log requests and responses for compliance needs.

• Analytics and Trends: Identify patterns and optimize your AI workflows using the ingested data.

---

Error Handling

Ensure proper error handling in your integration:

• 401 Unauthorized: Verify your API key is correct and has access.

• 404 Not Found: Check if the endpoint URL is correct.

• 422 Unprocessable Entity: Validate the request body for missing or invalid fields.

• 500 Internal Server Error: Retry or contact support if the issue persists.

---

Next Steps

Get into advanced analytics with Requesty’s Insight Explorer.