Getting Started
Quickstart
Start tracking AI agent costs in under 5 minutes. No changes to your existing AI calls needed — the SDK wraps your client transparently.
1
Get an API key
Go to Settings → API Keys and create a new key. Copy it — you won't be able to see it again.
AGENTRICS_API_KEY=agntcs_sk_...2
Install the SDK
Python
pip install agent-cost-optimizerJavaScript / TypeScript
npm install @agentrics/sdk3
Wrap your AI client
Python
import os
from anthropic import Anthropic
from agent_cost_optimizer import Agentrics
agentrics = Agentrics(api_key=os.environ["AGENTRICS_API_KEY"])
# Wrap your existing client — no other changes needed
client = agentrics.wrap_anthropic(
Anthropic(),
agent_id="support-classifier",
)
# Use exactly as before
response = client.messages.create(
model="claude-3-haiku-20240307",
max_tokens=256,
messages=[{"role": "user", "content": "Classify this ticket..."}],
)
# Cost is tracked automaticallyJavaScript / TypeScript
import OpenAI from "openai";
import { Agentrics } from "@agentrics/sdk";
const agentrics = new Agentrics({ apiKey: process.env.AGENTRICS_API_KEY! });
// Wrap your existing client
const client = agentrics.wrapOpenAI(new OpenAI(), "research-agent");
// Use exactly as before
const response = await client.chat.completions.create({
model: "gpt-4o-mini",
messages: [{ role: "user", content: "Summarize this document..." }],
});
// Cost is tracked automatically4
See your data
Open the Dashboard — costs appear in real time as your agents run. The spending breakdown shows per-agent and per-model costs, and the recommendations engine flags opportunities like model downgrades and call batching.
Next steps
- Python SDK reference → — all constructor options and wrapper methods
- JavaScript SDK reference → — TypeScript types and usage patterns
- REST API reference → — post tracking events directly without the SDK