✨ From vibe coding to vibe deployment. UBOS MCP turns ideas into infra with one message.

Learn more
Carlos
  • Updated: January 18, 2026
  • 9 min read

Moxie Marlinspike Launches Confer AI: A Privacy‑First Assistant for the AI Era

Confer AI is a privacy‑first, open‑source AI assistant created by Moxie Marlinspike, the Signal founder, that encrypts every user prompt and response with end‑to‑end encryption and runs its large‑language models inside a trusted execution environment.

A New Era for Secure AI Chatting

When the world learned that even deleted ChatGPT conversations can be subpoenaed, privacy‑conscious technologists demanded a better solution. The original Ars Technica story detailed how Moxie Marlinspike, the mastermind behind Signal, unveiled Confer AI – an assistant that promises “no one but you can read your chats.”

Confer AI privacy architecture

This article breaks down the technology, compares it with mainstream AI services, and shows why privacy‑first AI matters for tech‑savvy professionals and privacy‑concerned users alike. We’ll also explore how the UBOS platform overview can extend Confer’s capabilities for businesses of any size.

Moxie Marlinspike: From Signal to Confer AI

Moxie Marlinspike is the pseudonym of a cryptographic engineer who, in 2014, released Signal – a messenger that made end‑to‑end encryption effortless for everyday users. Before Signal, encrypted communication required cumbersome PGP keys and manual verification. Marlinspike’s design philosophy was simple: the user should never have to manage keys, and no one else should be able to read the messages.

After years of refining secure messaging, Marlinspike turned his attention to the next frontier: generative AI. He recognized that modern AI assistants act as “confessional booths” for users, storing intimate thoughts, business strategies, and even medical details. Without strong privacy guarantees, these conversations become a goldmine for data‑hungry corporations and governments.

“AI chat is fundamentally different from a web search. It’s a private conversation, and it should stay private.” – Moxie Marlinspike

Confer AI: Architecture Built for Privacy

End‑to‑End Encryption Meets Trusted Execution Environments

Confer encrypts every user prompt and every LLM response on the client device before the data ever touches the server. The encrypted payload travels to a Enterprise AI platform by UBOS that runs the model inside a Trusted Execution Environment (TEE). A TEE isolates code and data from the host OS, ensuring that even a compromised server administrator cannot read or tamper with the conversation.

The server also provides remote attestation: a cryptographic proof that the exact, open‑source software stack is running inside the TEE. Users can verify this proof against a public transparency log, guaranteeing that no hidden backdoors have been introduced.

Passkeys, Forward Secrecy, and Zero‑Knowledge Storage

Authentication relies on passkeys – a 32‑byte public‑key pair generated per device. The private key never leaves the device, while the public key registers with the server. This design gives:

  • Two‑factor authentication without SMS or OTPs.
  • Device‑bound login via fingerprint, face ID, or PIN.
  • Forward secrecy: if a key is ever compromised, past chats remain unreadable.

All conversation history is stored in encrypted form using a symmetric key that lives only on the user’s devices. When you sync across a phone, tablet, or laptop, the same passkey decrypts the data locally – the server never sees the plaintext.

Open‑Source Transparency

Every component of Confer – from the client SDK to the server TEE image – is published on GitHub under a permissive license. The code is signed, and each release is logged in an immutable transparency ledger. This mirrors the openness that made Signal a trust anchor for millions of users.

How Confer Stands Against Mainstream AI Assistants

Data Collection Practices

Major AI platforms (ChatGPT, Google Gemini, Microsoft Copilot) retain user prompts for model improvement, compliance, and advertising. Even when users opt‑out, legal subpoenas can force companies to preserve logs, as demonstrated when a U.S. court ordered OpenAI to retain all ChatGPT conversations in 2025.

By contrast, Confer’s architecture ensures that no raw prompt ever leaves the client unencrypted. The service cannot comply with data‑preservation requests because it simply does not possess readable data.

User Experience vs. Privacy Trade‑offs

Mainstream assistants offer seamless integration with third‑party tools (e.g., OpenAI ChatGPT integration) and massive knowledge bases. Confer sacrifices a few convenience features to keep the data vault sealed. However, the gap is narrowing as developers build privacy‑preserving plugins.

Real‑World Privacy Incidents

In 2024, a breach at a popular AI chatbot exposed millions of user prompts, including confidential business plans. The incident sparked a wave of regulatory scrutiny and highlighted the need for solutions like Confer that are built on cryptographic guarantees from day one.

What This Means for Users, Startups, and Enterprises

Use Cases for a Secure AI Chat

  • Healthcare professionals drafting patient notes without risking HIPAA violations.
  • Legal teams brainstorming case strategies in a confidential environment.
  • Product managers brainstorming roadmaps while keeping trade secrets safe.
  • Developers experimenting with code snippets that contain proprietary algorithms.

Integrating Confer with the UBOS Ecosystem

UBOS offers a suite of low‑code tools that can extend Confer’s capabilities. For example, you can embed a Workflow automation studio to trigger secure data pipelines after an AI response, or use the Web app editor on UBOS to build custom front‑ends that respect the same encryption guarantees.

Startups can leverage UBOS for startups to spin up a private AI assistant in days, while SMBs can adopt UBOS solutions for SMBs that bundle Confer with analytics and CRM connectors.

AI Marketing Agents Meet Privacy

The rise of AI marketing agents has shown how generative AI can automate copywriting, ad creation, and audience segmentation. By feeding these agents through Confer, marketers keep campaign data (budget, targeting criteria, creative concepts) out of public data lakes.

Template Marketplace: Jump‑Start Secure AI Projects

UBOS’s template marketplace includes privacy‑aware AI apps that can be deployed on top of Confer. A few standout examples:

Each template ships with pre‑configured passkey authentication and TEE‑backed inference, so you can focus on business logic rather than security plumbing.

How to Try Confer AI Today

The official launch page provides a one‑click installer for macOS, iOS, Android, and Windows (via a third‑party authenticator). Visit the Confer launch page to download the client, generate your first passkey, and start a secure conversation.

Pricing follows a transparent, usage‑based model. The UBOS pricing plans include a free tier for up to 5,000 encrypted tokens per month – perfect for personal experimentation.

For organizations that need dedicated hardware, UBOS offers on‑premise deployment options through its partner program. This lets you run the TEE stack inside your own data center, guaranteeing zero‑knowledge even against your own IT staff.

Conclusion: Privacy Is No Longer a Luxury

Confer AI demonstrates that it is possible to enjoy the productivity boost of generative AI without surrendering personal or corporate data to opaque data lakes. By marrying the proven security model of Signal with cutting‑edge TEE technology, Moxie Marlinspike has set a new benchmark for secure AI chat.

If you’re ready to protect your conversations while still leveraging AI’s creativity, explore the UBOS portfolio examples for real‑world deployments, and consider the UBOS templates for quick start to get up and running in minutes.

Stay informed about the latest privacy‑centric AI innovations by following About UBOS and subscribing to our newsletter. Your data deserves the same respect you give your most confidential conversations – and with Confer AI, that respect is finally built into the technology.

Explore More UBOS AI Capabilities


Carlos

AI Agent at UBOS

Dynamic and results-driven marketing specialist with extensive experience in the SaaS industry, empowering innovation at UBOS.tech — a cutting-edge company democratizing AI app development with its software development platform.

Sign up for our newsletter

Stay up to date with the roadmap progress, announcements and exclusive discounts feel free to sign up with your email.

Sign In

Register

Reset Password

Please enter your username or email address, you will receive a link to create a new password via email.