Anuma launches private AI with one memory across every leading model

anuma-launches-private-ai-with-one-memory-across-every-leading-model
Anuma launches private AI with one memory across every leading model

The privacy-first AI app from Basic Attention Token co-creator Ankur Nandwani opens to the public with one subscription, access to every leading model, and one encrypted memory across them all

SAN FRANCISCO, April 27, 2026 (GLOBE NEWSWIRE)Anuma today opened to the public with one subscription for ChatGPT, Claude, Gemini, Grok, DeepSeek, and other leading AI models, plus a user-owned encrypted memory that carries context across all of them. More than 10,000 people used the product during its beta.

AI has gone mainstream faster than trust, portability, or data ownership have caught up. Stanford HAI’s 2026 AI Index found generative AI hit 53% adoption in three years, faster than the PC or the internet. Meanwhile, Reuters reported this month that lawyers are warning users their chatbot conversations can be used against them in court, and The Economist reported that Anthropic and OpenAI are increasingly locking access to their latest models into proprietary ecosystems. The result is more power in users’ hands and less control over what they give up to use it.

Anuma is built to replace that trade-off. One subscription gives users access to the world’s leading AI models in one place. One encrypted memory carries context across them. And that memory stays under the user’s control, stored locally on their device and separated from model inference.

Instead of starting over in every new chat, users can move between ChatGPT, Claude, Gemini, Grok, DeepSeek, Kimi, Qwen, Gemma, and other models without losing context. Instead of handing over personal history to whichever model they happen to be using that day, users keep their memory encrypted, user-owned, and out of model training.

“Today’s AI products want to trap your context inside their own ecosystem,” said Ankur Nandwani, Core Contributor of Anuma. “We built Anuma so your memory stays yours, no matter which model you use.”

How Anuma Works

Anuma is built on three core principles:

  • One memory across all of them. Preferences, projects, writing style, goals, files, and conversation history carry across every session and every supported model.
  • Every model in one app. Users can switch between leading models mid-conversation without losing context or maintaining separate subscriptions. Council Mode compares answers from several models side by side.
  • Private by default. By default, user data stays on the user’s device, with sensitive content encrypted before it is stored. Anuma cannot decrypt user content, even with full server access.

What Users Can Do

  • Chat with any leading AI model from one interface
  • Text AI over SMS and iMessage with no app download needed
  • Create images, videos, and audio with Creative Studio
  • Build working apps from a prompt or description (coming soon)
  • Solve real-world tasks such as billing, paperwork, and housing issues with Anuma agents (coming soon)

What sets Anuma apart is its memory architecture. Preferences, goals, project context, files, and conversation history live inside an encrypted vault on the user’s device. That memory loads automatically into conversations across models, so users do not have to repeat themselves or rebuild context from scratch every time they switch tools. Users can also review, edit, delete, import, or export their vault whenever they want.

Anuma is available now on the web, with iOS and Android coming soon. Users can also access AI by text message through SMS and iMessage.

A private-first approach to AI

Nandwani previously co-created Basic Attention Token, the token that powers Brave’s privacy-first browser economy. Brave’s transparency reporting shows 112.2 million monthly active users as of March 31, 2026. That scale matters because it proves that privacy is not a niche preference when the product is strong enough.

Anuma brings that same default-private posture to AI: user-owned memory, encryption on-device, support for open-source models where appropriate, and the ability to move across models without surrendering context to any single provider.

“Brave was created because the browser had become a surveillance tool and users had no real say in it,” Nandwani said. “AI is heading in the same direction. A lot of people want the benefits of the best models without giving up ownership of their memory, and Anuma gives them that choice.”

Pricing

Anuma’s free tier includes 100 credits per month and access to every major feature, including encrypted memory, multi-model chat, AI texting, and Anuma Agents. No credit card is required.

Starter is $9.99 per month. Pro is $19.99 per month.

About Anuma

Anuma is the privacy-first, all-in-one AI platform with persistent memory across every model, plus tools to chat, text, create, build, and solve. Your memory is encrypted on your device, never stored on Anuma’s servers, and never used to train any model. One subscription unlocks ChatGPT, Claude, Gemini, Grok, DeepSeek, Kimi, Llama, Mistral, Qwen, Gemma, and more. Anuma is powered by ZetaChain. Learn more at anuma.ai.

Press contact

Jonathan Covey

jonathan@anuma.ai

A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/e15ae2ad-f35e-4c48-9db2-09086b8811a1