Getting started

Welcome to the Aeon documentation!

This section will help you start tracking your AI Agent conversations in 5 minutes.

1. Get your credentials

1. Go to your dashboard
2. Copy your API key and Project ID

2. Install the SDK

Python:

bash
1pip install aeonpy

JavaScript/TypeScript:

bash
1npm install @aeon-ai/aeonjs

3. Initialize the Aeon client

python
1from aeon import Aeon
2
3# Use env vars for secrets
4aeon = Aeon(api_key="YOUR_API", project_id=123)

Or in JavaScript/TypeScript:

javascript
1import Aeon from "@aeon-ai/aeonjs";
2
3const aeon = new Aeon("API_KEY", 123);

4. Track conversations

After initializing the Aeon client, you can track conversations using the track_conversation method:

python
1aeon.track_conversation(
2    conversation_id="123",
3    provider="openai",
4    model="gpt-4",
5    messages=[
6        {
7            "role": "user",
8            "content": "...",
9            "timestamp": time.time(),
10        },
11        {
12            "role": "assistant",
13            "content": "...",
14            "timestamp": time.time(),
15        },
16    ]
17)

JavaScript / TypeScript

Use the following code to track conversations:

javascript
1await aeon.trackConversation({
2  conversationId: "abc123",
3  provider: "openai",
4  model: "gpt-4",
5  messages: [
6    {
7      role: "user",
8      content: "...",
9      timestamp: Date.now(),
10    },
11    {
12      role: "assistant",
13      content: "...",
14      timestamp: Date.now(),
15    },
16  ],
17});

Aeon automatically gives you insights like cost breakdowns, user frustrations, feature requests, and more.

You can see all tracked conversations in your dashboard.

Endpoint

The Conversations API allows you to send user conversations to Aeon for analysis and tracking.

No code tools

To integrate Aeon with no code tools like n8n or Copilot Studio, you can use the Conversations API endpoint directly.

Request

The request body should be a JSON object with the following structure:

json
1{
2  "conversation_id": "<string>",
3  "provider": "<string>",
4  "model": "<string>",
5  "messages": [
6    {
7      "role": "user",
8      "content": "<string>",
9      "timestamp": 1763259080
10    },
11    {
12      "role": "assistant",
13      "content": "<string>",
14      "timestamp": 1763259082
15    }
16  ]
17}

Parameters

• conversation_id (str): Unique identifier for the conversation. If called again with the same ID, the conversation will be updated.
• provider (str): Name of the LLM provider (e.g., "openai", "anthropic").
• model (str): Model used for the conversation (used to calculate cost).
• messages (List[Dict]): List of message objects, each with role, content, and timestamp.

Example (curl)

Send a conversation payload using curl:

bash
1curl --request POST \
2  --url https://withaeon.com/api/v1/projects/123/agents/conversations \
3  --header 'Authorization: Bearer <token>' \
4  --header 'Content-Type: application/json' \
5  --data '{
6        "conversation_id": "123",
7        "provider": "openai",
8        "model": "gpt-4",
9        "messages": [
10        {
11            "role": "user",
12            "content": "...",
13            "timestamp": 1763259080
14        },
15        {
16            "role": "assistant",
17            "content": "...",
18            "timestamp": 1763259082
19        }
20        ]
21    }'

Next Steps

• Explore your dashboard to see your tracked conversations.