Completions

The Completions endpoint creates text given a prompt. Use it for classic instruction following and text generation when you do not need multi-turn message structures.

---

Create completion

POST https://api.aifoundryhub.com/v1/completions

Creates a completion for the provided prompt and parameters.

---

Example request

curl

curl -X POST "https://api.aifoundryhub.com/v1/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AI_FOUNDRY_HUB_API_KEY" \
  -d '{
    "model": "babbage-002",
    "prompt": "Write a haiku about dawn",
    "max_tokens": 64,
    "temperature": 0.7
  }'

node.js

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.AI_FOUNDRY_HUB_API_KEY,
  baseURL: "https://api.aifoundryhub.com/v1",
});

const resp = await client.completions.create({
  model: "babbage-002",
  prompt: "Write a haiku about dawn",
  max_tokens: 64,
  temperature: 0.7,
});

console.log(resp.choices[0].text);

go

package main

import (
  "context"
  "fmt"
  "os"

  openai "github.com/openai/openai-go"
  "github.com/openai/openai-go/option"
)

func main() {
  client := openai.NewClient(
    option.WithAPIKey(os.Getenv("AI_FOUNDRY_HUB_API_KEY")),
    option.WithBaseURL("https://api.aifoundryhub.com/v1"),
  )

  params := openai.CompletionNewParams{
    Model: "babbage-002",
    Prompt: openai.CompletionNewParamsPromptUnion{
      OfString: openai.String("Write a haiku about dawn"),
    },
    MaxTokens:   openai.Int(64),
    Temperature: openai.Float(0.7),
  }

  ctx := context.Background()
  resp, err := client.Completions.New(ctx, params)
  if err != nil { panic(err) }
  fmt.Println(resp.Choices[0].Text)
}

python

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.getenv("AI_FOUNDRY_HUB_API_KEY"),
    base_url="https://api.aifoundryhub.com/v1",
)

resp = client.completions.create(
    model="babbage-002",
    prompt="Write a haiku about dawn",
    max_tokens=64,
    temperature=0.7,
)

print(resp.choices[0].text)

Returns

• When stream is false: a text completion object.
• When stream is true: a Server-Sent Events stream of incremental completion chunks.

SSE snippet

data: {"id":"cmpl_...","object":"text_completion.chunk","choices":[{"index":0,"delta":{"text":"The"},"finish_reason":null}]}

data: {"id":"cmpl_...","object":"text_completion.chunk","choices":[{"index":0,"delta":{"text":" dawn"},"finish_reason":null}]}

data: [DONE]

Example response (non-streaming)

{
  "id": "cmpl-abc123",
  "object": "text_completion",
  "created": 1714569952,
  "model": "babbage-002",
  "choices": [
    {
      "index": 0,
      "text": "Pale light spills over\nWaking earth in gentle hush\nDay breaks soft and new",
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 6,
    "completion_tokens": 24,
    "total_tokens": 30
  }
}

The completion object

---

Streaming chunk (delta) object

Note: The stream ends with a terminal data: [DONE] line. Some providers also include a final non-chunk summary event with usage.

Streaming examples

curl

curl -N -X POST "https://api.aifoundryhub.com/v1/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AI_FOUNDRY_HUB_API_KEY" \
  -d '{
    "model": "babbage-002",
    "prompt": "Write a haiku about dawn",
    "stream": true,
    "max_tokens": 64
  }'

node.js

const stream = await client.completions.create({
  model: "babbage-002",
  prompt: "Write a haiku about dawn",
  stream: true,
  max_tokens: 64,
});

for await (const part of stream) {
  const delta = part?.choices?.[0]?.delta?.text;
  const legacy = part?.choices?.[0]?.text;
  process.stdout.write(delta ?? legacy ?? "");
}

go

st := client.Completions.NewStreaming(ctx, params)
defer st.Close()
for st.Next() {
  chunk := st.Current()
  for _, ch := range chunk.Choices {
    if ch.Delta != nil {
      fmt.Print(ch.Delta.Text)
    } else {
      fmt.Print(ch.Text)
    }
  }
}
if err := st.Err(); err != nil { panic(err) }

python

stream = client.completions.create(
    model="babbage-002",
    prompt="Write a haiku about dawn",
    stream=True,
    max_tokens=64,
)

for event in stream:
    # Prefer delta.text if present, else fall back to text
    delta_text = getattr(getattr(event.choices[0], "delta", None), "text", None)
    text = delta_text or getattr(event.choices[0], "text", "")
    print(text, end="")