# Gridline

> Gridline is the API for training your own AI models and running them in batch: fine-tuning, distillation and batch inference. Text fine-tuning and batch are live today: sign up and add credits by card to start.

One API for fine-tuning, distillation and batch inference.

## Keep the client you already use

Point OpenAI's Python client at Gridline and your first fine-tune is three calls away.

```python
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.gridlinegpu.com/openai/v1",
    api_key=os.environ["GRIDLINE_TOKEN"],
)

training = client.files.create(file=open("train.jsonl", "rb"), purpose="fine-tune")
job = client.fine_tuning.jobs.create(model="qwen2.5-7b-instruct", training_file=training.id)
print(job.gridline["quote"])  # an estimate, before the job runs
```

## For agents

- One index: llms.txt lists every public page and the docs, with the quickstart inline.
- Every page as Markdown: Add .md to a page's URL, or send Accept: text/markdown to the URL you already have.
- An API it already knows: Files, fine-tuning jobs and batches follow OpenAI's API. Change the base URL and the key.
- The home page's Markdown is https://gridlinegpu.com/index.md. A Markdown answer ends with the URL of the page it was made from.
- The API base URL is https://api.gridlinegpu.com/openai/v1. The docs keep the token in the GRIDLINE_TOKEN environment variable; never print or commit it.
- Fine-tunes are quoted before they run, and every job is paid from prepaid credits. Ask the person before creating a fine-tune or a batch.

## Pages

- [Gridline](https://gridlinegpu.com/?mode=machine)
- [Docs](https://gridlinegpu.com/docs)
- [Pricing](https://gridlinegpu.com/pricing?mode=machine)
- [Get started](https://gridlinegpu.com/signup)
- [Privacy](https://gridlinegpu.com/privacy)
- [Terms](https://gridlinegpu.com/terms)
- [Refunds](https://gridlinegpu.com/refund)

## Docs

- [Fine-tuning](https://gridlinegpu.com/docs/fine-tuning)
- [Batch API](https://gridlinegpu.com/docs/batch-api)
- [Distillation](https://gridlinegpu.com/docs/distillation)
- [Models](https://gridlinegpu.com/docs/models)
- [Migrate from OpenAI](https://gridlinegpu.com/docs/openai-migration)
- [Advanced: your own script](https://gridlinegpu.com/docs/own-script)

## Optional

- [llms-full.txt](https://gridlinegpu.com/llms-full.txt)