OpenAI-compatible API
Open-weight models, one simple API.
Zither Tech runs leading open-weight language models on cloud GPUs and serves them through an OpenAI-compatible endpoint. Change one line of code and pay only for the tokens you use.
from openai import OpenAI
client = OpenAI(
base_url="https://api.zithertech.com/v1",
api_key="YOUR_API_KEY",
)
resp = client.chat.completions.create(
model="qwen3.8-27b",
messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)Why Zither Tech
Open models without the infrastructure work.
Drop-in compatible
Keep the OpenAI SDKs and tools you already use. Point them at our base URL with your Zither Tech key and your existing code keeps working.
Pay per token
Add prepaid credits and pay only for the tokens you process, at the rates listed below. No subscriptions, seats or minimum spend.
Private by default
We deploy and run every model ourselves on cloud GPUs. We never train on your prompts or outputs, and we don't store them.
Models & pricing
Simple per-token pricing.
Prices in US dollars per million tokens. Input and output tokens are billed separately.
| Model | Developer | Context | Input | Output |
|---|---|---|---|---|
| Qwen3.8 27Bqwen3.8-27b | Qwen | 64K tokens | $0.35 | $2.00 |
Usage is deducted from your prepaid balance after each request. The console shows the token counts and cost of every call. Online payment by card is coming soon.
Get started
From sign-up to first response in minutes.
Add credits, create a key
Top up your balance (online payment is coming soon), then create an API key. Each key can be limited by model, spend and expiry date.
curl https://api.zithertech.com/v1/chat/completions \
-H "Authorization: Bearer $ZITHER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "qwen3.8-27b", "messages": [{"role": "user", "content": "Hello!"}]}'FAQ
Common questions.
Which SDKs and tools work with Zither Tech?
Anything that lets you set a custom OpenAI base URL: the official OpenAI libraries for Python and Node.js, frameworks such as LangChain and LlamaIndex, and many chat apps and editor plugins. Use https://api.zithertech.com/v1 as the base URL.
How does billing work?
You buy prepaid credits in US dollars. Each request deducts its input and output tokens at the listed rates. There are no monthly fees, and the console shows the cost of every request. Card payments through Stripe are coming soon.
Do credits expire? Can I get a refund?
Credits are valid for 12 months from purchase. Unused credits can be refunded within 30 days of purchase; email support to request one. Our Terms of Service have the details.
Do you train on my data?
No. We don't use your prompts or outputs to train or improve any model, and we don't store their content. Our Privacy Policy explains what we do keep, such as token counts for billing.
What can't I use the API for?
Illegal, harmful or abusive uses are not allowed. See the Acceptable Use Policy for the full list.
How do I get help?
Email support@zithertech.com. We aim to reply within one business day.
Start building with open models.
Create an account and get your API key.