End-to-end encrypted history
Only you can read your chat history. Your chats are encrypted with a key derived from your password before they ever reach our servers. Not even we can read them.
Private AI chat, built in Switzerland
Veila gives you today's most capable language models without handing over your identity. Your prompts reach the providers anonymously, and your history stays end-to-end encrypted.
Why it matters
We share thoughts, plans, worries and work with AI. That makes chat logs some of the most personal data there is — and privacy matters more than ever.
Health questions, relationship advice, business ideas, half-finished drafts. AI assistants see it all.
Many services log your prompts, link them to your identity, and use them for training — or worse, sell the data.
We forward your prompts anonymously, store your chats end-to-end encrypted, and never log your identity.
How it works
One simple idea: everyone's prompts leave through the same door, so no one can tell who asked what.
Send a message from the Veila chat app, just like any other AI assistant.
Our servers remove identifying metadata like your IP address and route only the message onward. Nothing is logged.
The provider sees one request among thousands from Veila — and answers without knowing who you are.
The usual way
What the provider sees:
With Veila
What the provider sees:
Features
Only you can read your chat history. Your chats are encrypted with a key derived from your password before they ever reach our servers. Not even we can read them.
We proxy your prompts, so model providers can't build a profile of you.
Hosted under some of the world's strongest privacy laws.
Talk to models from OpenAI, Anthropic, Google, Mistral and more — even mid-chat.
Use folders and favorites to keep track of your work.
Propose and vote on what we build next. You're in control.
Pricing
Use the free tier daily, or pay only for what you use. No subscription required.
$0
A daily token allowance to get you going.
from $10
Top up credits, pay only for what you use.
Model costs plus a transparent 30% markup — that's it. See the pricing documentation for how the free tier and credits work.
FAQ
Can't find what you're looking for? Read the documentation or write to hey@veila.ai.
When you use ChatGPT directly, OpenAI knows exactly who you are, when you chat, and can build a detailed profile of your interests and behavior. With Veila, we strip away all identifying information before forwarding your prompts, so the AI providers only see an anonymous request from our servers — not you. See also our provider comparison or why to trust us.
No, Veila is an independent product and is not affiliated with OpenAI, Anthropic, or other AI model providers. We provide a custom privacy-focused interface that connects to these AI services while protecting your identity and data.
While we integrate with models from various providers, we operate independently and our privacy protections are our own implementation, not provided by the AI companies themselves.
We store your chat history encrypted end-to-end so only you can read it. We also store your account credentials and usage for billing. We never log your prompts, IP addresses, or any metadata that could identify your conversations. Your chats pass through our servers in plain text for AI processing, but are immediately discarded — never stored or analyzed.
Important: Third-party AI providers (like OpenAI) receive your prompts and may process them according to their own data policies. You must comply with their usage policies.
When you send a prompt, our servers remove all identifying metadata (your IP, user agent, timestamps, etc.) and forward only the message content to the AI provider. From their perspective, all requests come from our servers, not individual users. It's like having thousands of people wear identical masks — the AI can respond to the message, but can't tell who asked.
Veila offers a curated selection of leading models from OpenAI, Anthropic, Google, Mistral, Meta, DeepSeek, Qwen and others — from fast, affordable everyday models to the most capable frontier models. You can switch between models at any time, even mid-conversation, to compare responses or use the best model for each task. We regularly add new models as they are released.
All your encrypted chat data is permanently deleted from our servers. Since we don't log prompts or store identifying information about your conversations, there's nothing else to remove. Your data is truly gone.
No email, phone number, or real name required. Just create a username and password to start chatting for free. For paid features, we only need payment information — no identity verification or personal details.
Free users help us achieve better anonymization through volume, while paid users cover costs for everyone. We keep pricing simple: just a modest 30% markup on model costs, which supports ongoing development and ensures strong privacy protections. No ads, no data selling, no hidden agenda.
The 30% markup covers essential operational costs that enable our privacy-focused service:
Unlike subscription services with hidden costs, our transparent markup means you only pay for what you actually use, plus a small amount to keep the service running and improving. For model details, visit our provider comparison and our pricing documentation.
Your chat history is encrypted end-to-end using industry-standard AES-256 encryption. Your password-derived key encrypts everything client-side before it reaches our servers, so we literally cannot read your stored chats even if we wanted to. See also our detailed encryption documentation.
There's no enterprise plan yet, but we're interested in organizational use cases. If you think your company should use Veila, or if you want to give your employees access to privacy-focused AI chat, contact us at hey@veila.ai. We'd love to discuss your needs and explore how Veila can benefit your team.
Yes! Developer API access is coming soon. You'll be able to integrate Veila's anonymized AI access into your applications, giving your users the same privacy benefits without the complexity of managing multiple AI provider relationships. Please upvote this in our feature board if you're interested.
Since your chats are encrypted end-to-end, a server breach would only expose encrypted data that's useless without your password. We're building data export tools so you can download your encrypted chats anytime. Our Swiss hosting and strong privacy laws provide additional protection against forced shutdowns.
Yes, you must comply with all third-party AI provider policies. This includes OpenAI's Usage Policies which prohibit harmful, illegal, or inappropriate content. Violations may result in immediate account termination.
AI Content Disclaimer: AI responses may be inaccurate, biased, or inappropriate. You are responsible for reviewing and validating all AI-generated content before use. We are not liable for any harm caused by AI outputs.
About us
We're privacy nerds, engineers, and AI users tired of being the product. Veila is our response — a tool that respects your data, by design. Read the story on our About page.
Independent audit planned as we scale.
No personal info needed. Use it free. Pay only if you love it.
Get started now