Token Factory · Documentation

Omniva Token Factory Documentation

Token Factory is easy to use with an OpenAI-compatible /v1 API that works with your existing workflows. During Early Access please contact sales for help getting started.

Make your first call API reference

Endpoint

https://api.tokenfactory.omniva.com/v1

Env var

OMNIVA_API_KEY

Example model

Omniva/Kimi-K2.6

New here

First time? Make your first call.

Migrating from OpenAI

Swap base_url and key, keep the rest.

Operating in prod

Watch latency, spend, and errors live.

First call

Request

from openai import OpenAI

client = OpenAI(
    api_key="sk-...",
    base_url="https://api.tokenfactory.omniva.com/v1",
)

resp = client.chat.completions.create(
    model="Omniva/Kimi-K2.6",
    messages=[{"role": "user", "content": "Say hi."}],
)
print(resp.choices[0].message.content)

200application/json

{
  "id": "chatcmpl-7b1f6c…",
  "object": "chat.completion",
  "model": "Omniva/Kimi-K2.6",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hi! How can I help?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 9, "completion_tokens": 7, "total_tokens": 16 }
}

The same OpenAI SDK works — point base_url at https://api.tokenfactory.omniva.com/v1.Open the full Quickstart

Build something

Quickstart

Your first chat call

from openai import OpenAI

client = OpenAI(base_url=BASE_URL, api_key=KEY)
resp = client.chat.completions.create(
    model="Omniva/Kimi-K2.6",
    messages=[{"role": "user", "content": "Hi"}],
)

PY/v1/chat/completions

Open recipe

Streaming

Stream tokens to the browser

const stream = await client.chat.completions.create({
  model: "Omniva/Kimi-K2.6",
  messages,
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0].delta.content ?? "");
}

TS/v1/chat/completions

Open recipe

Embeddings

Embed and search documents

res = client.embeddings.create(
    model="<embedding-model-id>",
    input=["the cat sat on the mat"],
)
vec = res.data[0].embedding
print(len(vec), vec[:3])

PY/v1/embeddings

Open recipe

Migration

Move from OpenAI in one line

# Before: client = OpenAI(api_key=OPENAI_KEY)
# After:
client = OpenAI(
    base_url="https://api.tokenfactory.omniva.com/v1",
    api_key=AISTUDIO_KEY,
)
# Everything else stays the same.

Open recipe

Auth

Create and rotate API keys

# Create the key in the dashboard:
#   Developer → API Keys → Create
export OMNIVA_API_KEY="sk-oais..."
echo "key set (${OMNIVA_API_KEY:0:7}...)"

Open recipe

Errors

Handle failures cleanly

try:
    client.chat.completions.create(model=MODEL, messages=msgs)
except openai.RateLimitError as e:
    retry_after = e.response.headers.get("Retry-After", 30)
    time.sleep(int(retry_after))

Open recipe