Oxlo Vision Backend

The AI gateway behind Oxlo Vision: unified, fast, and developer-friendly.

Tip

What is this? This backend centralizes access to multiple state-of-the-art AI models (LLMs, vision, audio, image generation, and embeddings) under a single API. Built with Micronaut and Java 21.

---

Project Vision

Building with many AI providers is usually messy: different payloads, different endpoints, and different model conventions. Oxlo Vision Backend solves that with one integration layer.

It acts as an API Gateway + Catalog:

1. Integrated Catalog (/api/ias): discover available models by category and retrieve integration metadata.
2. Universal Proxy (/v1/*): send standardized requests while the backend routes traffic to the corresponding upstream model on Oxlo.ai.

---

Architecture and Endpoints

The backend is split into two main domains.

1) AI Catalog (/api/ias)

Discovery services for frontend and external clients.

Method	Endpoint	Description
GET	/api/ias	Returns categories and consolidated model catalog.
GET	/api/ias/{category}/{model}	Returns model metadata, route info, and cURL sample.

Supported categories: chat, reasoning, coding, vision, image-gen, audio, embedding.

2) Universal Proxy (/v1)

Execution layer that forwards standardized calls to model-specific upstream APIs.

Method	Endpoint	Typical workloads
POST	/v1/chat/completions	Chat and reasoning models
POST	/v1/detect	Object detection models
POST	/v1/images/generations	Image generation models
POST	/v1/audio/speech	Text-to-speech models
POST	/v1/audio/transcriptions	Speech-to-text models (multipart/form-data)
POST	/v1/embeddings	Embedding/vector models

---

Quickstart

Requirements:

• JDK 21

1. Build

git clone <repo-url>
cd back-end
./mvnw clean install

2. Run in Dev Mode

./mvnw mn:run

Backend URL: http://localhost:8080

---

Consumption Examples

Chat completion

curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama-3.3-70b",
    "messages": [
      {"role": "user", "content": "Summarize this technical document."}
    ]
  }'

Image generation

curl http://localhost:8080/v1/images/generations \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stable-diffusion-1.5",
    "prompt": "An AI robot winning a hackathon",
    "size": "1024x1024"
  }'

Audio transcription

curl http://localhost:8080/v1/audio/transcriptions \
  -F "file=@sample.mp3" \
  -F "model=whisper-large-v3"

---

Built for OxBuild with Micronaut, Java 21, and Oxlo.ai.