Documentation

Get started in 5 simple steps

Setting up your Oryxen AI account is quick and straightforward. Follow these steps to generate your API key and start using our confidential computing platform.

1

Generate your API key

On the login page, click "Generate a new API key" to begin your secure setup.

Step 1: Generate API key
2

Understand why security matters

Your mnemonic phrase and API key are the foundation of your account security—learn why they matter.

Step 2: Security lecture
3

Secure your mnemonic

Receive your unique mnemonic phrase and a reminder to store it somewhere safe—this is your master backup.

Step 3: Save mnemonic
4

Get your API key

Your API key is revealed for the first and only time—copy it immediately and store it securely.

Step 4: API key displayed
5

Activate your account

Select your preferred plan, complete the payment, and your API key is instantly activated and ready to use.

Step 5: Payment and activation

Logging In & Accessing Oryxen AI

Log in securely and start using Oryxen AI through our web chat or API.

Web Chat

Use our secure, end-to-end encrypted web interface. No setup required—just sign in and start chatting.

Launch Web Chat →

API Integration

Build your own interface using our REST API. Full OpenAI-compatible endpoints with confidential computing guarantees.

View API Docs →

Logging In

1

Go to the Oryxen login page: app.oryxen.ai/login.html

2

Enter your API key in the text box

3

Click Connect

4

Your messages are automatically encrypted

Browser Requirements

Chrome 104+ | Firefox 132+ | Safari 17.4+ | Edge 104+

If your browser is too old, you will see a warning and cannot log in.

The Sidebar

The sidebar on the left shows key areas:

Model Selector

A dropdown at the top lets you choose which AI model to use. Most of the models support text chat and documents analysis. Many have limited support for image analysis and optical character recognition (OCR). Vision models specifically are optimized for image/video analysis.

Model TypeHow to IdentifyWhat You Can Do
Vision Name contains "vl" or "vision" Attach images 🖼️
Chat Everything else Regular text chat and documents analysis

Tip: If you want to send images, switch to a vision model first!

Token Usage

Two stat cards show your usage:

This Chat

Input and output tokens for the current conversation

Tokens

Remaining / Used Today / Monthly Used

Running low? Contact your administrator to add more tokens.

Using the Chat

The Chat Input Area

At the bottom of the chat area you will find:

📄 report.pdf 📊 data.csv ← document chips (RAG)
🖼️ img.jpg ← image thumbnails
Type a message...
image attach
document attach
send button
🖼️ Attach Image Send images to vision-capable models
📚 Attach Document Attach files for RAG-powered answers
Send Send your message

Header Buttons

At the top of the chat area, the header contains:

ButtonWhat It Does
+ New Chat Clears messages, documents, and images
Export Downloads the full chat as a Markdown file
Logout Ends your session and clears encryption keys

Sending a Message

1

Type your question in the text box (auto-expands up to 200px)

2

Press Enter or click the ➜ send button

3

The AI responds in the chat window

Message Features

  • Assistant responses support Markdown, code blocks, tables, and math (LaTeX)
  • Click the copy button on any completed assistant message to copy the raw text
  • While the AI is thinking, you will see a spinner with "Oryxen is thinking" — no copy button appears during loading

⏰ Session Duration

Your session stays alive for up to 4 hours of inactivity. Any request to the server restarts the 4-hour clock, so you stay logged in while you keep working.

Attaching an Image (Vision Models)

You can send images to AI models that support vision:

1

Select a vision model from the dropdown (e.g., qwen2vl)

2

Click the 🖼️ image button next to the text box

3

Select an image file (JPG, PNG, GIF, WebP, etc.)

4

The image appears as a thumbnail above the text box

5

Type your question about the image and send

Why is the 🖼️ image button disabled?

If the 🖼️ image button is grayed out, hover over it — you will see:

"🎨 Switch to a vision model to attach images"

Select a model with vision capabilities from the dropdown.

Attaching Documents (RAG)

RAG = Retrieval-Augmented Generation

Attach documents and ask questions about their contents. The AI reads your files and bases its answers only on what you uploaded.

How to Attach a Document

1

Click the 📚 document button (books icon) next to the text box

2

Select one or more documents

Supported Formats

CategoryFormats
📃 OfficePDF, Word (DOCX), PowerPoint (PPTX), Excel (XLSX)
📝 TextTXT, Markdown (MD), CSV, JSON, XML, YAML
📊 OpenDocumentODT, ODS, ODP
🌐 WebHTML, RTF
3

Each document appears as a chip above the text box

Document Chip States

StateAppearanceMeaning
🔄 ProcessingTeal dotFile is uploading and being embedded
✅ ReadyClean chipAvailable for RAG context
❌ FailedRed textUpload error — try again or remove
4

RAG is automatically enabled when at least one document is ready

5

Type your question and send ➜

How it works: You only need to type your question. The AI sees your attached documents' content as background context and answers based on them. If the answer is not in your documents, the AI will say so clearly.

Removing Documents

ActionHow
Remove oneClick the × on a document chip
Delete from serverImmediate — no undo
Select multiple filesHold Ctrl or Shift when clicking 📚

Privacy: Documents are private — only you can see them. They are isolated by your API key and deleted when you remove them.

Starting a Fresh Conversation

Click + New Chat in the header to clear everything:

New Chat clears:

  • All messages
  • All attached documents
  • All attached images

You get a completely clean slate.

Also remember that your session automatically expires after 4 hours of inactivity — the server keeps you logged in while you work.

Support & Recovery

Recovering a Lost Key

If you saved your 12-word backup phrase but lost your API key:

1

Go to the Oryxen login page

2

Click the "Recover Key" tab

3

Enter your 12-word phrase exactly as written (order matters!)

4

Click Recover Key

5

Your API key appears — copy and store it safely

Critical: If you lost both your API key and your 12-word phrase, your account cannot be recovered.

This is by design — nobody, including administrators, can access your key or phrase.

Security Tips

Do This

  • Keep your API key secret
  • Write down your 12-word phrase on paper
  • Log out when done (especially on shared computers)
  • Keep your browser updated

Never Do This

  • Share your API key
  • Email your phrase
  • Screenshot your phrase
  • Reuse your key on other sites

Troubleshooting

"Invalid API key"

Possible CauseFix
Missing oryxen_ prefixCopy the entire key, including oryxen_
Key not activatedContact your administrator — key may still be PENDING

"Encryption unavailable" warning

Fix:

  • Update your browser
  • Try a different browser

Details:

Chrome 104+ / Firefox 132+ / Safari 17.4+

Chrome, Firefox, or Safari recommended

Note: The browser must support WebCrypto API, we rely on this for safe cryptographic operations.

"No documents found" with RAG

CheckAction
Still processing?Wait for the teal dot to disappear (Ready)
Failed upload?Remove the chip and try again
Vague question?Use words that actually appear in the document

"No embedding model available" tooltip

The document button is disabled because no provider with embedding models is configured.

Contact your administrator to add an embedding-capable provider, such as:

  • Internal provider with embedding-model
  • Any OpenAI-compatible provider with an embeddings endpoint

Image attach button is disabled

CauseFix
Non-vision model selectedSwitch to a vision model (e.g., qwen2vl)
Hover shows tooltip"🎨 Switch to a vision model to attach images"

Page keeps redirecting to login

1

Clear your browser cookies for the Oryxen website

2

Try logging in again with your API key

3

If it persists, contact us

Getting Help

Issue TypeWho to Contact
Account issues (activation, lost keys)Contact us
Technical problemsCheck browser console (press F12) for error messages

OryxenWeb v0.20.27 | OryxenProxy v0.18.5 | Tinia v0.8.5

User Guide

Download User Guide

Get the complete PDF for offline reading, detailed walkthroughs, and full screenshots.

Download

Read User Guide Online

Browse the interactive HTML version with clickable table of contents and easy navigation.

Open Online Guide

Develop Quick Start

Get your first API call working in under 2 minutes. Oryxen provides an OpenAI-compatible API processed inside hardware-isolated Trusted Execution Environments (TEEs).

1

Base URL: https://api.oryxen.ai/v1

2

Authentication: Send your API key as Bearer <your-api-key> in the Authorization header

3

Copy-paste this cURL into your terminal:

curl https://api.oryxen.ai/v1/chat/completions \
  -H "Authorization: Bearer your-api-key-here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "oryxen-llm",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Need More?

  • Jump to Code Examples for Python, JavaScript, Go, Rust, and more
  • Read the full API Guide for detailed endpoint documentation

Code Examples

Python
JavaScript
cURL
Go
Rust
Haskell
Julia
import requests

API_KEY = "your-api-key-here"
BASE_URL = "https://api.oryxen.ai/v1"

response = requests.post(
    f"{BASE_URL}/chat/completions",
    headers={
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json"
    },
    json={
        "model": "oryxen-llm",
        "messages": [
            {"role": "user", "content": "Hello, how are you?"}
        ]
    }
)

print(response.json()["choices"][0]["message"]["content"])

API Guide

Read API Guide Online

Browse the interactive HTML version with full endpoint documentation and integration examples.

Open API Guide
×