Legacy Completions

Endpoint

POST https://api.bve.me/v1/completions

Requires Authorization: Bearer sk-bve-YOUR_KEY.

Emulated endpoint

Fuelix does not support /v1/completions natively. BVE Gateway emulates it by converting the request to a chat completions call internally and translating the response back to the completions format. The upstream is always /chat/completions.

Limitations

• best_of and logprobs are not supported by the emulation layer and return 400 unsupported_parameter.
• Array prompts are joined with \n\n before being sent as a single user message.
• The emulated id is derived from the chat completion id (chatcmpl- → cmpl-).

Request body

{
  "model": "gpt-4o",
  "prompt": "The capital of France is",
  "max_tokens": 10
}

Field	Type	Required	Description
model	string	Yes	Model ID
prompt	string | string[]	No	Prompt text. Arrays are joined with \n\n.
max_tokens	integer	No	Max tokens to generate
temperature	number	No	Sampling temperature (0–2)
top_p	number	No	Nucleus sampling (0–1)
n	integer	No	Number of completion choices to generate; must be ≥ 1
best_of	integer	No	Must be ≥ 1; always rejected — not supported by the emulation layer (returns 400 unsupported_parameter)
echo	boolean	No	Prepend the normalized prompt text to each returned choice
logprobs	integer	No	Top-N log probabilities; must be ≥ 0. Always rejected — not supported by the emulation layer (returns 400 unsupported_parameter)
stop	string | string[]	No	Up to 4 stop sequences
presence_penalty	number	No	Presence penalty (−2 to 2)
frequency_penalty	number	No	Frequency penalty (−2 to 2)
user	string	No	End-user identifier (max 256 chars)
seed	integer	No	Deterministic seed
stream	boolean	No	Enable streaming SSE mode. Chunks use object: "text_completion" format (same as non-streaming but delivered as SSE).

Response

{
  "id": "cmpl-abc123",
  "object": "text_completion",
  "created": 1716288000,
  "model": "gpt-4o",
  "choices": [
    {
      "text": " Paris",
      "index": 0,
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 7,
    "completion_tokens": 2,
    "total_tokens": 9
  }
}

Response headers

BVE Gateway adds the following headers to every /v1/completions response:

Header	Example	Description
X-Request-Id	550e8400-…	Server-generated UUID for this request
X-BVE-Client-Id	my-trace-123	Echo of the client-supplied X-Request-Id (when present and valid: alphanumeric + -_., ≤ 128 chars). Absent when the client did not send X-Request-Id or the value failed validation.
X-BVE-Latency	87	Total gateway latency in milliseconds
X-BVE-Model	gpt-4o	Model ID for this request
X-BVE-Key-Name	prod-key	Display name of the authenticated key

The standard X-RateLimit-* per-key headers (requests per minute, per day, monthly caps, token limits) are also present. See Rate Limits & Quotas for the full table and example output.

cURL example

curl https://api.bve.me/v1/completions \
  -H "Authorization: Bearer sk-bve-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "prompt": "The capital of France is",
    "max_tokens": 5
  }'

SDK example

TypeScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-bve-YOUR_KEY',
  baseURL: 'https://api.bve.me/v1',
});

const completion = await client.completions.create({
  model: 'gpt-4o',
  prompt: 'The capital of France is',
  max_tokens: 5,
});

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

Python

from openai import OpenAI

client = OpenAI(
    api_key="sk-bve-YOUR_KEY",
    base_url="https://api.bve.me/v1",
)

completion = client.completions.create(
    model="gpt-4o",
    prompt="The capital of France is",
    max_tokens=5,
)

print(completion.choices[0].text)

Array prompts

Array prompts are joined with \n\n before being forwarded as a single user message.

TypeScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-bve-YOUR_KEY',
  baseURL: 'https://api.bve.me/v1',
});

const completion = await client.completions.create({
  model: 'gpt-4o',
  prompt: ['Translate to French:', 'Hello, world!'],
  max_tokens: 20,
});

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

Python

from openai import OpenAI

client = OpenAI(
    api_key="sk-bve-YOUR_KEY",
    base_url="https://api.bve.me/v1",
)

completion = client.completions.create(
    model="gpt-4o",
    prompt=["Translate to French:", "Hello, world!"],
    max_tokens=20,
)

print(completion.choices[0].text)

Echo mode

When echo: true is set, the normalized prompt text is prepended to each choice’s text field.

TypeScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-bve-YOUR_KEY',
  baseURL: 'https://api.bve.me/v1',
});

const completion = await client.completions.create({
  model: 'gpt-4o',
  prompt: 'The capital of France is',
  max_tokens: 5,
  echo: true,
});

// text includes the prompt: "The capital of France is Paris"
console.log(completion.choices[0].text);

Python

from openai import OpenAI

client = OpenAI(
    api_key="sk-bve-YOUR_KEY",
    base_url="https://api.bve.me/v1",
)

completion = client.completions.create(
    model="gpt-4o",
    prompt="The capital of France is",
    max_tokens=5,
    echo=True,
)

# text includes the prompt: "The capital of France is Paris"
print(completion.choices[0].text)

Streaming

Streaming is supported. Set "stream": true and BVE Gateway will transform the upstream chat.completion.chunk SSE events into text_completion SSE events in real time.

curl https://api.bve.me/v1/completions \
  -H "Authorization: Bearer sk-bve-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-4o", "prompt": "The capital of France is", "max_tokens": 10, "stream": true}'

Each SSE chunk uses object: "text_completion" with a choices[].text field instead of choices[].delta.content:

data: {"id":"cmpl-abc","object":"text_completion","created":1716288000,"model":"gpt-4o","choices":[{"text":" Paris","index":0,"logprobs":null,"finish_reason":null}]}

data: {"id":"cmpl-abc","object":"text_completion","created":1716288000,"model":"gpt-4o","choices":[{"text":"","index":0,"logprobs":null,"finish_reason":"stop"}]}

data: [DONE]

TypeScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-bve-YOUR_KEY',
  baseURL: 'https://api.bve.me/v1',
});

const stream = await client.completions.create({
  model: 'gpt-4o',
  prompt: 'The capital of France is',
  max_tokens: 10,
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.text ?? '');
}

Python

from openai import OpenAI

client = OpenAI(
    api_key="sk-bve-YOUR_KEY",
    base_url="https://api.bve.me/v1",
)

stream = client.completions.create(
    model="gpt-4o",
    prompt="The capital of France is",
    max_tokens=10,
    stream=True,
)

for chunk in stream:
    print(chunk.choices[0].text or "", end="", flush=True)

Gateway validation

The gateway validates required fields before forwarding to the emulation path.

Required fields

Missing / invalid	Code	param
model absent	missing_required_parameter	"model"
model not a string	invalid_type	"model"
prompt not a string or array	invalid_type	"prompt"
stream not a boolean	invalid_type	"stream"

Unsupported parameters

Condition	Code	param
best_of not a positive integer	invalid_value	"best_of"
best_of set (any valid integer)	unsupported_parameter	"best_of"
logprobs not a non-negative integer	invalid_value	"logprobs"
logprobs set (any valid integer)	unsupported_parameter	"logprobs"

Optional field constraints

Parameter	Invalid condition	Code	param
temperature	Not a number	invalid_type	"temperature"
temperature	Outside [0, 2]	invalid_value	"temperature"
top_p	Not a number	invalid_type	"top_p"
top_p	Outside [0, 1]	invalid_value	"top_p"
max_tokens	Not a number (e.g. a string)	invalid_type	"max_tokens"
max_tokens	Not a positive integer (float or ≤ 0)	invalid_value	"max_tokens"
n	Not a number (e.g. a string)	invalid_type	"n"
n	Not a positive integer (float or ≤ 0)	invalid_value	"n"
echo	Not a boolean	invalid_type	"echo"
presence_penalty	Not a number	invalid_type	"presence_penalty"
presence_penalty	Outside [-2, 2]	invalid_value	"presence_penalty"
frequency_penalty	Not a number	invalid_type	"frequency_penalty"
frequency_penalty	Outside [-2, 2]	invalid_value	"frequency_penalty"
stop	Not a string or array	invalid_type	"stop"
stop	Array with more than 4 elements	invalid_value	"stop"
stop[N]	Not a string	invalid_type	"stop[N]"
user	Not a string	invalid_type	"user"
user	Longer than 256 characters	invalid_value	"user"
seed	Not a number (e.g. a string)	invalid_type	"seed"
seed	A float (e.g. 1.5)	invalid_value	"seed"

All 400 responses use the standard error envelope:

{
  "error": {
    "message": "temperature must be between 0 and 2",
    "type": "invalid_request_error",
    "param": "temperature",
    "code": "invalid_value"
  }
}

Next steps

Chat Completions Recommended replacement — richer features: tool calls, multimodal content, logprobs

Rate Limits & Quotas Per-key RPM/RPD/monthly enforcement and Retry-After handling

Errors Full error code reference with envelope shape and handling examples