Move a call from the Anthropic SDK
A Model answers on the OpenAI chat API, so a call made with the Anthropic SDK moves to the OpenAI SDK. The system prompt goes away: Facet writes and keeps the Model’s Instructions.
Python
Before, with Claude:
import anthropic
client = anthropic.Anthropic()
msg = client.messages.create(
model="claude-sonnet-5-5",
max_tokens=512,
system="Label each ticket billing, bug or account...",
messages=[{"role": "user", "content": ticket}],
)
label = msg.content[0].text
After, with a Facet Model:
import os
from openai import OpenAI
client = OpenAI(base_url=os.environ["FACET_BASE_URL"], api_key=os.environ["FACET_API_KEY"])
reply = client.chat.completions.create(
model="support-triage",
messages=[{"role": "user", "content": ticket}],
)
label = reply.choices[0].message.content
TypeScript
import OpenAI from "openai";
const client = new OpenAI({ baseURL: process.env.FACET_BASE_URL, apiKey: process.env.FACET_API_KEY });
const reply = await client.chat.completions.create({
model: "support-triage",
messages: [{ role: "user", content: ticket }],
});
const label = reply.choices[0].message.content;
What changes
| Anthropic SDK | Facet Model |
|---|---|
system |
Left out; Facet sends the Model’s Instructions |
msg.content[0].text |
reply.choices[0].message.content |
max_tokens required |
Optional |
stop_reason |
finish_reason |
To make the Model from the call, paste the old system prompt and a few real inputs into your coding agent and ask it to use Facet, or open New model in the web app to connect one. See Use Facet from an agent.
Next steps
- Call a Model: every parameter and header a call takes
- Return structured output: get answers as fields or JSON