Everything you need to know about Anuma. Can’t find what you’re looking for? Say hello
What Anuma is, how it works, and how it compares to other AI platforms.
Anuma is a privacy-focused, all-in-one AI platform with unified, on-device memory and context across app, web, and SMS or iMessage. Access leading models like ChatGPT, Claude, and Gemini, alongside private open-source models such as DeepSeek and Kimi, all in one interface. Compare up to four models at once, create images and videos, solve real-world tasks with specialized AI agents, while maintaining full ownership of your data.
ChatGPT is a model. Claude is a model. Anuma is the platform that brings all of them together. One subscription gets you every major AI model, a unified memory that works across all of them, and full control over your data. You pick the best model for each task. Anuma makes sure it already knows you.
Memory is what Anuma knows about you. Context is what the AI sees in a conversation. Anuma turns your memory into context automatically, so every model you use already knows your background, preferences, and goals from the first message.
Yes. You can move between models within Anuma without restarting. Your unified memory and conversation context carry over.
No. A single Anuma subscription gives you access to every model on the platform. There is no need to pay for separate ChatGPT, Claude, or Gemini plans.
How Anuma remembers your preferences, context, and history across every model.
Anuma memory is a private, persistent knowledge base you build about yourself. It loads automatically into every AI conversation so the AI always knows your context without you repeating it.
ChatGPT memory only works inside ChatGPT. Anuma memory works across every model including Claude, Gemini, DeepSeek, and more. You also own the data and can export it anytime.
Yes. You can add, edit, or delete any memory entry at any time. You can also set individual entries to private or standard. Deletion is permanent.
Yes. You can mark any memory entry as private so it stays on your device and is never sent to an AI model. You control exactly what gets shared and what stays hidden.
Yes. Anuma memory works with ChatGPT, Claude, Gemini, DeepSeek, and every other model available on the platform. Your memory travels with you no matter which model you use.
Your memory is encrypted and stored on your device. Private memory entries never leave your device at all.
You can export your full memory at any time from your account settings.
Deleted content is permanently removed. Anuma doesn't keep hidden copies for ads or training. Deleted means deleted.
Your encrypted memory is tied to your secure identity, not a single device. When you sign in on a new device, your encrypted memory can be restored safely without exposing it to anyone else.
Memory Engine automatically indexes and searches past conversations. My Memory stores specific facts and preferences that are explicitly saved.
Every message and memory is converted into a vector that captures its meaning. Search finds the most similar vectors so you don't need exact keyword matches.
My Memory uses supersession detection. When a newer memory is similar to an older one and enough time has passed, the newer version is automatically prioritized.
There is no hard limit. My Memory holds thousands of entries for fast access.
How Anuma manages context across models, devices, and conversations.
Your preferences, conversation history, and saved facts travel with you across models and devices. Start on mobile, continue on desktop. Switch from Claude to GPT. Your context follows.
Anuma selects the most relevant memories and recent conversation history for each request. You can also manually control what the AI sees.
Yes. Your memory and context carry over when you switch models. The new model picks up where the last one left off.
Anuma manages context automatically. Older parts of long conversations are summarized so the most important information stays accessible.
No. Your memory loads into every new conversation automatically. You never start from zero.
How Anuma encrypts your data, separates open and closed-source memories, and keeps you in control.
Anuma uses 256-bit AES-GCM encryption. Keys are derived from your wallet signature using SHA-256 and HKDF with domain separation. Encrypted values are version-prefixed so encrypted and plaintext data coexist seamlessly. Encryption happens on your device before data goes anywhere.
No. By default, all data stays on your device. If you opt in to encrypted cloud backup, only encrypted blobs are stored on Cloudflare R2. Anuma cannot read them.
No. Anuma never holds your encryption keys. Data is encrypted on your device before it leaves. Even with full server access, employees cannot decrypt your content.
When you use standard AI models, your message is sent to their servers so they can generate a response. Anuma limits what they see and keeps your private memory separate. While external providers may log data under their own policies, they don't get permanent access to your full history or vault.
Open-source models do not retain your data or use it for training. Closed-source models may retain data according to their own policies. Because of this, open-source models can access all your memories. Closed-source models can only access memories from closed-source conversations unless you explicitly share more through My Memory.
No. Anuma never trains on your memory data, sells it, or shares it with third parties. Your memory exists solely to serve you.
Local data can be cleared from the app at any time. If you enabled cloud backup, you can delete all encrypted backups from settings. Deletion is immediate and permanent.
Yes. Images, audio, and other content you create in Creative Studio are stored on your device and not shared with anyone.
Using multiple AI models, switching between them, and comparing with Council Mode.
Closed-source: ChatGPT, Claude, Gemini, and Grok. Open-source: DeepSeek, Kimi, MiniMax, Cerebras Qwen 3, and GLM. All are included on every plan, including the free tier. New models are added as they become available.
It depends on your task. Claude is strong for analysis and long documents. ChatGPT is versatile for general use and writing. Gemini is great for real-time information. DeepSeek excels at coding and math. You can use Council Mode to run multiple models at once and compare their answers.
Yes. You can switch to a different model at any point in a conversation. Anuma rebuilds the full context window for the new model using the same memory, conversation summaries, and history. No information is lost.
Open-source models have publicly available code and weights, and process your data with zero retention. Closed-source models are proprietary and run on provider servers. They generally deliver the highest capability for creative and reasoning tasks. On Anuma, you can switch between them in one tap.
Council Mode sends the same prompt to multiple AI models at once. You see every response side by side and choose the one that works best. It's the fastest way to compare how different models handle the same question.
You can run up to four models in a single Council Mode prompt. Pick any combination from the models available on Anuma.
No. Each model has different strengths. Some are better at creative writing, others at coding or research. Anuma lets you access all of them to pick the right tool for each task.
Yes. Use open source models when privacy matters most. Use closed source models for enhanced capability. Switch when it makes sense for the task.
Open source models have advanced significantly. DeepSeek and others perform strongly for everyday tasks, and they are ideal when data protection is the priority.
Yes. Anuma regularly adds new models as they become available. Your memory works with every model automatically so there is nothing to set up when a new one is added.
Understanding the difference between open and closed source AI models on Anuma.
An open source LLM is an AI model whose code and weights are publicly available. This means the model can be audited, customized, and run independently. On Anuma, open source models do not retain your data or use it for training.
A closed source LLM is a proprietary AI model run by a company like OpenAI, Anthropic, or Google. They offer high performance but the code is not public. Data may be retained according to the provider's policies.
Anuma offers DeepSeek, Kimi, Qwen, and more. New open source models are added regularly.
Anuma offers Claude, GPT, Gemini, Grok, and more from leading AI providers.
Open source models do not retain your data or use it for training. Combined with Anuma's on device encryption, this provides the highest level of privacy.
Creative Studio, app generation, AI agents, Texting AI, and more.
Creative Studio is Anuma's built-in suite for generating images, videos, and audio. It brings together leading AI models from multiple providers -- including Flux, Nano Banana, Veo, Sora, Kling, and Gemini Audio -- so you can create everything from one app with one subscription.
Yes. Audio Studio supports both music generation and sound effects. Describe a mood or style for music, or describe any sound for effects. You can control duration, toggle looping, and adjust prompt influence.
Every image, video, and audio clip you generate is automatically saved to your Anuma Library. You can search, organize, re-download, or share from one place.
A capability inside Anuma that generates interactive mini-apps from a single prompt. Describe what you want and get a working interface you can use immediately -- budget trackers, workout timers, quiz generators, project dashboards, simple games, and more.
No. Just describe what you want in plain language. You can iterate by saying things like "add a dark mode" or "make the chart weekly."
AI agents that go beyond answering questions. They analyze documents, generate letters, send emails on your behalf, and guide you through real-world processes like lease disputes and money recovery.
Purpose-built AI assistants that you reach by text message. Each agent specializes in one domain such as fitness, legal, health, or finance. Save a phone number and text it like any other contact -- no app download required.
You can describe any app you need and Anuma will build it. Internal tools, dashboards, calculators, trackers, and more. No coding required.
Yes. You can share apps you build with others via a link.
Anuma's real world agents are currently being built. Join the waitlist to get early access when they launch.
Plans, credits, upgrades, and how Anuma compares to paying for each model separately.
Yes. Anuma Free gives you 100 credits per month and access to every feature, including all AI models, all modes, encrypted memory, AI Friends, and Anuma Agents. No credit card required.
Credits are used each time you send a message, generate an image, or use voice and video features. Different models use different amounts. You can upgrade to a higher tier anytime.
Every plan includes access to all AI models, all modes (Private, Standard, Council), encrypted memory, AI Friends, text, image, voice, video, and app generation, and Anuma Agents. Paid plans unlock higher usage limits and more storage.
ChatGPT Plus is $20/month. Claude Pro is $20/month. Gemini Advanced is $20/month. That is $60/month for three separate tools with no shared memory. Anuma Pro is $19.99/month and includes all three plus unified memory, Private Mode, Council Mode, and more.
Council Mode is included in all Anuma plans. Each model response counts toward your usage, so running three models uses roughly three times what a single model would.
No. Credits reset at the start of each billing cycle. If you need more credits during a cycle, you can upgrade to a higher tier.
You can still access your chats, memory, and saved files. To send new messages or generate content, you can wait for your next monthly reset, or upgrade to a higher plan.
Yes. You can upgrade, downgrade, or cancel from your account settings at any time. Changes take effect at the start of your next billing cycle. No fees, no penalties.
Yes. Encrypted memory is included on every plan, including Free. Your data is never used for model training. Anuma does not sell or share your information.
Yes. Upgrade, downgrade, or cancel from your account settings at any time. No cancellation fees. Changes take effect at the start of your next billing cycle.
Devices, apps, SMS, web, and how to access Anuma everywhere.
Anuma is available as a web app, an iOS app, and an Android app. You can also access AI through SMS and iMessage via Texting AI Agents.
Yes. Council Mode works on Anuma's web app, iOS app, and Android app. Same experience across all devices.
Yes. Your memory and conversation history sync across all devices in real time. Start on mobile, continue on desktop. Everything is encrypted and stored securely.
Wallet login means you never hand over your email, phone number, or any personal information. It's the most private way to use AI. You can also use a regular account if you prefer.
Earn by sharing Anuma with your audience.
Anyone can apply. Bloggers, creators, developers, and AI enthusiasts are all welcome.
Commissions are paid monthly. You earn a percentage of each referred subscription for as long as the user remains subscribed.
No. There is no cap on affiliate earnings. The more people you refer, the more you earn.
Reach out anytime. We're happy to help.