# OpenFPGA

> FPGA-accelerated AI inference. OpenAI-compatible API. Drop-in replacement — just change the base URL.

## API

OpenFPGA provides an OpenAI-compatible chat completions API backed by FPGA hardware acceleration. No proprietary SDK required — use the standard OpenAI client libraries.

- Base URL: https://api.openfpga.ai/v1
- Auth: Bearer token (API key prefix: ofpga_sk_live_)
- Model: llama-3.1-8b-fpga (Llama 3.1 8B Instruct on FPGA)

## Docs

- [API Reference](https://app.openfpga.ai/#/docs)
- [OpenAPI Spec](https://app.openfpga.ai/.well-known/openapi.yaml)
- [Full Context (llms-full.txt)](https://app.openfpga.ai/llms-full.txt)
- [Quickstart](https://app.openfpga.ai/#/quickstart)

## API Endpoints

- [Chat Completions](https://docs.openfpga.ai/api/chat-completions): POST /v1/chat/completions — chat inference
- [Embeddings](https://docs.openfpga.ai/api/embeddings): POST /v1/embeddings — text embeddings
- [Models](https://docs.openfpga.ai/api/models): GET /v1/models — list available models

## Quick Start

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.openfpga.ai/v1",
    api_key="ofpga_sk_live_..."
)

response = client.chat.completions.create(
    model="llama-3.1-8b-fpga",
    messages=[{"role": "user", "content": "Hello!"}]
)
```