You run it on your laptop. Now ship it.

BitsRun turns any open model you use with Ollama or Hugging Face into your own private API. It scales to zero when nobody is calling it, and you pay by the second only while it runs.

Deploy your first model See prices

From $0.95 per GPU hour, billed per second. Prepaid, top up from $10.

Terminalqwen3-8b, NVIDIA L4

$ bitsrun deploy ollama:qwen3:8b

Not deployed$0.000264 per second

$0.00000

0.0 s billed

From local model to private API in three steps

Step 1
Pick the model you already run
An Ollama name like qwen3:8b, a Hugging Face repo, or your own fine-tune as GGUF or safetensors.
Step 2
Deploy it with one command
bitsrun deploy gives you a private endpoint on a GPU that fits the model. No Dockerfile, no server to manage.
Step 3
Call it from your app
Use the OpenAI client you already have. Change the base URL, keep your code.

Pay for the minutes you use, not the month

Most apps are busy for a few hours a day. Move the slider to see what that costs compared with keeping the same GPU running all month.

BitsRun, scales to zero

$85.54 / month

3 h a day for 30 days at $0.95 per hour

Same GPU running all month

$693.79

You keep 88% of that bill.

Keep your code. Change one line.

Endpoints speak the OpenAI API, with streaming, and GGUF models also speak Ollama's API. Point your existing client at your BitsRun URL and add your key.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://01k0d4xq8m6v2t9r5w3b7n1c4e.endpoint.bitsrun.com/v1",
    api_key=os.environ["BITSRUN_API_KEY"],
)
reply = client.chat.completions.create(
    model="qwen3:8b",
    messages=[{"role": "user", "content": "Hello"}],
)

bitsrun, in your terminal

Run open AI models as private APIs, and code with any of them right here in your terminal. One command deploys, manages keys and opens a coding agent that works with your BitsRun endpoints or any other provider.

Install bitsrun

$ bitsrun deploy ollama:qwen3-coder:30b
$ bitsrun
▍ add a health check route to the express app

How billing works

Startup is never billed: The meter starts when your model is ready to answer and stops when your idle timeout ends.
No surprise bills: You buy credit up front, and we email you when it runs low. At a zero balance your endpoints stop, so you never owe more than you paid.
Your prompts stay yours: We never log prompts or outputs, and we never train on them. Each endpoint is private to your account.
Requests wait, they do not fail: If your endpoint is waking up, requests are held until it is ready instead of returning errors.

Questions, answered

Which models can I run?

Large language models in GGUF (through llama.cpp) or safetensors (through vLLM), plus embedding models. Image and audio models are next on the roadmap.

How long does a cold start take?

On a warm GPU, a model is usually ready in a few seconds to about 40 seconds. When a new GPU has to start, GGUF and embedding models take about 2 to 3 minutes, and safetensors models on vLLM about 3 to 4 minutes. None of that time is billed, and your requests wait for it (up to 8 minutes) instead of failing.

What does the idle timeout do?

It is how long your endpoint stays up after the last request, so the next request answers instantly. You choose it, from 60 seconds to 60 minutes, and you pay per second while it waits.

Can I keep an endpoint always on?

Yes. Turn on always-on and your endpoint keeps a dedicated GPU that never sleeps, so there are no cold starts. It is billed by the second while it is on, at a monthly rate: $1,180 a month for an L4.

Do I need a GPU or Docker?

No. Most text and embedding models on Ollama or Hugging Face deploy by name, and you never write a Dockerfile. A few kinds are not supported yet: see models that cannot be deployed.

Ship the model you already trust

Deploy your first model