Connection Registry

The connection registry lets application code register preconfigured clients or connection settings by name. A .prompty file can then use connection.kind: reference without embedding secrets in the file.

Register and look up a connection

Python

from openai import AzureOpenAI
import prompty

client = AzureOpenAI(
    azure_endpoint=os.environ["AZURE_ENDPOINT"],
    api_key=os.environ["AZURE_API_KEY"],
)

prompty.register_connection("my-foundry", client=client)
client = prompty.get_connection("my-foundry")
prompty.clear_connections()

TypeScript

import { registerConnection, getConnection, clearConnections } from "@prompty/core";
import AzureOpenAI from "openai";

const client = new AzureOpenAI({
  endpoint: process.env.AZURE_ENDPOINT,
  apiKey: process.env.AZURE_API_KEY,
});

registerConnection("my-foundry", client);
const conn = getConnection("my-foundry");
clearConnections();

C#

using Prompty.Core;

ConnectionRegistry.Register("my-foundry", client);
var conn = ConnectionRegistry.Get("my-foundry");
ConnectionRegistry.Clear();

Rust

use prompty;
use serde_json::json;

prompty::register_connection("my-foundry", json!({
    "kind": "key",
    "endpoint": std::env::var("AZURE_ENDPOINT")?,
    "apiKey": std::env::var("AZURE_API_KEY")?,
}));

let conn = prompty::get_connection("my-foundry");
prompty::clear_connections();

Reference from .prompty

model:
  id: gpt-4o
  provider: foundry
  connection:
    kind: reference
    name: my-foundry

Use references for environment-specific values, authenticated SDK clients, and test fixtures. Use explicit key, remote, anonymous, foundry, or oauth connections when the configuration should live directly in the prompt.

For connection data shapes, see Connection.