Skip to main content

​
Getting Started with RedPill

Set up your private AI assistant in minutes. No credit card required for the free tier.

​
Create Your Account

1

Sign Up

Go to chatgpt.redpill.ai and create your account with email or OAuth (Google, GitHub).
2

Verify Email

Check your inbox and click the verification link.
3

Access Chat

You’ll be redirected to the chat interface automatically.

​
Your First Conversation

Once you’re in, you’ll see the chat interface:
  1. Select a model from the dropdown (DeepSeek V3 is a great default)
  2. Type your message in the input box
  3. Press Enter or click Send
That’s it - your message is already end-to-end encrypted.
All models in RedPill are Confidential AI models running in GPU TEE. Your data never leaves the secure enclave.

​
Understanding the Interface

​
Model Selector

Click the model name at the top to switch between available models:
ModelBest For
DeepSeek V3General conversations, reasoning
Qwen 2.5Multilingual, coding
GLM-4Chinese language, analysis
Llama 3.3Fast responses, chat

​
Chat Controls

  • New Chat - Start a fresh conversation
  • History - Access previous chats (encrypted locally)
  • Settings - Adjust preferences

​
Privacy Indicator

Look for the 🔒 lock icon - it confirms your connection is E2E encrypted.

​
Upload Documents (Private RAG)

Want to chat about your own files?
  1. Click the 📎 attachment icon
  2. Select files (PDF, DOCX, MD, TXT)
  3. Wait for processing (files are encrypted immediately)
  4. Ask questions about your documents
Files are processed entirely within the TEE. They’re encrypted before upload and never stored in plaintext.

​
Free Tier Limits

The free tier includes:
  • 10 messages per day
  • Access to basic Confidential AI models
  • E2E encryption (same as paid plans)
  • Single device
Need more? Upgrade to Pro for unlimited messages and premium features.

​
Tips for Getting Started

Clear, detailed prompts get better responses. Include context and specify the format you want.
Upload documents you reference often to your knowledge base - it’s more efficient than re-uploading.
Each model has strengths. DeepSeek excels at reasoning, Qwen at multilingual tasks. Experiment!
The 🔒 icon confirms E2E encryption is active. If you don’t see it, refresh the page.

​
Next Steps

Explore Models

Learn about 50+ Confidential AI models

Private RAG

Upload and query your documents privately

Pricing

Compare Free, Pro, and Enterprise plans

How Privacy Works

Deep dive into our privacy architecture