Veyllo API preview

OpenAI-compatible endpoints for chat, vision, and transcription. Point any OpenAI client at them and use the same shapes you already know, one key for everything.

POST /v1/chat/completions

Send messages, get a completion, streaming or not. Standard OpenAI Chat Completions, one model: veyllo-chat.

request

{
  "model": "veyllo-chat",
  "messages": [
    { "role": "user", "content": "Summarize this." }
  ]
}

Vision, same endpoint

Drop an image into any message and veyllo-chat reads it. No separate model, no extra setup.

request

{
  "model": "veyllo-chat",
  "messages": [{ "role": "user", "content": [
    { "type": "text", "text": "What's in this image?" },
    { "type": "image_url",
      "image_url": { "url": "data:image/jpeg;base64,…" } }
  ]}]
}

POST /v1/audio/transcriptions

Upload audio, get text back. Standard OpenAI transcription shape, metered per second, processed in the EU. One model: veyllo-transcribe. Pass language (de, en, …) for the best accuracy, or leave it out and the language is detected automatically.

request

curl https://api.veyllo.app/v1/audio/transcriptions \
  -H "Authorization: Bearer vaf_live_…" \
  -F file=@meeting.mp3 \
  -F model=veyllo-transcribe \
  -F language=de

{ "text": "Guten Morgen, fangen wir an…" }

Drop-in compatible

Already use the OpenAI SDK? Change two lines and you’re on Veyllo.

python

from openai import OpenAI

client = OpenAI(
    api_key="vaf_live_…",
    base_url="https://api.veyllo.app/v1",
)
client.chat.completions.create(
    model="veyllo-chat",
    messages=[{"role": "user", "content": "Hello!"}],
)
client.audio.transcriptions.create(
    model="veyllo-transcribe",
    file=open("meeting.mp3", "rb"),
)

Get an API key See pricing