Morse vs Mammouth: where your chats are kept, what providers keep, and what you pay
The Morse teamPublished Facts checked 11 min read
Morse vs Mammouth comes down to who holds your chats and who can read each prompt. As of November 2026, Mammouth, from the Paris company Mammouth AI SAS, sells Claude, GPT, Gemini, Mistral, Grok, and more from $10 a month billed yearly, stores your conversations on its own servers for as long as you subscribe, and sends each prompt to the company behind the model you picked. Morse keeps your chat history encrypted in your browser, not on its servers, and lets you choose per chat who reads the prompt, up to E2EE chats that only an enclave your browser has checked can open.
We read Mammouth's home page, its pricing page (monthly and yearly), its privacy policy and terms of service (both last updated November 13, 2026), the terms' data processing addendum, acceptable use policy, and sub-processor list, its "Privacy with model providers" page (last updated September 28, 2026), its quota policy, and its FAQ, all on November 20, 2026. We didn't open an account or send any prompts. Mammouth earns credit for publishing a page that quotes each AI company's training terms one by one, and a sub-processor list that says where each one works.
Morse is a private AI chat app with GPT, Claude, Gemini, DeepSeek, and uncensored models, encrypted so even Morse can't read your chats. Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse sends your message under its own account, so the model provider never learns your name, email, account, or IP address. Morse Pro is $19 a month, with unlimited messages on Pro models.
Morse vs Mammouth at a glance
As of November 20, 2026. Every Mammouth fact comes from the Mammouth pages listed under Sources at the end.
| Morse | Mammouth | |
|---|---|---|
| Company | Watermark Labs, Inc., San Francisco | Mammouth AI SAS, Paris, with contracts under French law and the courts of Paris |
| Where your prompt is read | Anonymous: by the model provider, without knowing it's you. Private: by a host that agreed to keep nothing. E2EE: only inside an enclave your browser verified | By the model's provider; the privacy policy names OpenAI, Anthropic, Mistral AI, Google, xAI, Perplexity, and OpenRouter among them |
| Who verifies it | In E2EE, your browser checks the enclave's hardware attestation before anything is sent | No check described; the policy says "Data in transit are encrypted." |
| Where chat history lives | Stored in your browser, encrypted, not on Morse's servers | On Mammouth's servers, hosted in Germany, kept "as long as your subscription is active" |
| What the model provider learns about you | Never your name, email, account, or IP address | No identity or Mammouth IDs, but possibly "a fully anonymised user ID" |
| Training | Morse doesn't train AI on your chats, and no AI provider trains on what you send through Morse | Mammouth says it doesn't, and that its providers don't train on prompts from Mammouth |
| Ads | None | None; the policy says it doesn't sell or share data for advertising |
| Models | 80+, open and closed, in one account, uncensored ones included on Pro and Max | 21 text and search models, 4 image models, and 3 video models in its home page Q&A |
| Free plan | Yes, no card needed | None on the pricing page |
| Paid plans | Pro $19 a month ($180 a year), Max $49 a month ($468 a year) | Starter $12, Standard $24, Expert $72 a month before VAT; $10, $20, and $60 a month billed yearly |
| How usage is counted | Pro: unlimited messages on Pro models, plus 250 monthly credits for the top models. Max: everything in Pro, plus 5,500 monthly credits | Quotas per 3-hour session; at a model's limit you move to a lighter one |
Where does Mammouth keep your chats?
On its own servers, for as long as you pay. Mammouth's privacy policy lists "your prompts, system interactions, and conversations" as user data, kept "for the display of past conversations in history," and says it is "Stored as long as your subscription is active." The sub-processor list in its terms names two main sub-processors, for computing and hosting, both working in Germany, and a separate provider in the EU holds uploaded files. Mammouth's privacy policy says personal data is processed and stored within the European Economic Area, with Standard Contractual Clauses for anything sent outside it, "e.g., for certain AI model providers."
Every retention period the privacy policy publishes, in one place:
| What Mammouth keeps | How long, in its privacy policy's words |
|---|---|
| Conversations and connected Google data | "Stored as long as your subscription is active" |
| Conversations you delete | "we might retain it for up to 30 rolling days for abuse detection" |
| Account data (name, email) | "Retained for up to 1 year after the end of your subscription" |
| Security data (usage logs, IP addresses, device information) | "Retained for up to 1 rolling year, then deleted automatically" |
| Payment data | "usually 5 to 7 years, depending on your country of residence" |
| Support emails | "Retained for up to 2 years from the date of your last interaction" |
You can delete your history "directly from your account settings," in the home page's words, and two Mammouth pages describe what follows a little differently. Mammouth's FAQ says "You can delete all your data at anytime. It will then be deleted from all our servers," while the privacy policy allows the 30-day abuse window above. The policy also says Mammouth may disclose data to a court or law-enforcement authority "when required by law, a court order, or other legal process," and since the chats sit on its servers, they're within reach of such an order. How courts have reached stored AI chats is in can AI chats be subpoenaed.
Do Mammouth's AI providers keep your prompts?
For up to 30 days, according to Mammouth's own provider page, though its home page says they keep none. The home page's privacy section reads "AI providers retain no data" and "Our providers do not retain prompts sent via Mammouth. No history is stored on their side." Mammouth's "Privacy with model providers" page is more exact. It lists two guarantees it enforces across providers, and the second is "No long-term retention: providers do not retain Mammouth prompts beyond 30 days, generally only to monitor service abuse via logs." Its note on Mistral says the same: "They only retain data for 30 days to monitor abuse."
That page also says what a provider receives: your prompt, "relevant conversation history, document extracts, and applicable custom instructions," without your identity or your Mammouth user, chat, team, project, or file IDs, though "they may receive a fully anonymised user ID" for abuse prevention. OpenRouter serves as a routing layer that "may transmit prompts and responses to downstream model providers," which Mammouth says it holds to the same no-training and no-retention rules. Our Morse vs OpenRouter page covers OpenRouter's own side.
Mammouth's sub-processor list gives each provider's location. Mistral AI processes in the EU; OpenAI, Anthropic, and Google in the "EU / USA"; and xAI, Perplexity, and OpenRouter in the USA. So a prompt to Grok leaves the EU even though your Mammouth history stays in it. What a zero retention promise does and doesn't cover is in what is zero data retention.
Does Mammouth train on your data?
No. Mammouth's provider page says "We don't train any internal models, and we never use your conversations to train anything," and its home page says your content "is never used to train models." The data processing addendum in its terms repeats it: "Mammouth shall not use Customer Personal Data to train artificial intelligence models." For each provider, the same page quotes the API or commercial terms Mammouth relies on, such as Anthropic's "may not train models on Customer Content from paid Services." Its privacy policy adds that Mammouth makes money "exclusively from the subscriptions you pay," and says "We do not sell or share your data for advertising purposes."
Morse doesn't train AI on your chats either, and its privacy policy puts the provider side in one sentence: no AI provider trains on what you send through Morse.
How much does Mammouth cost compared with Morse?
Mammouth starts lower: $12 a month before VAT for Starter, against $19 for Morse Pro. The "from $10/month" on Mammouth's home page is the yearly price. With Pay Yearly selected, its pricing page shows Starter at $10 a month with "$120 billed in advance," Standard at $20 ($240), and Expert at $60 ($720), all "(excl. VAT)." The pricing page lists no free plan, and Mammouth's terms say free trials may be offered only "Under specific circumstances."
| Plan | Price | What it includes |
|---|---|---|
| Mammouth Starter | $12 a month, or $120 a year, before VAT | All its models, web search with Perplexity, images and video, $2 of API credit, the base quota |
| Mammouth Standard | $24 a month, or $240 a year, before VAT | The same features at 3 times Starter's quota, $4 of API credit |
| Mammouth Expert | $72 a month, or $720 a year, before VAT | 10 times Starter's quota, $10 of API credit |
| Morse Free | $0, no card | Chats encrypted on your device, Anonymous and Private chats, web search, and chat tools |
| Morse Pro | $19 a month, or $180 a year | Unlimited messages on Pro models with your choice of model, 250 monthly credits for the top models, E2EE chats, uncensored models, deep research, memory, and plugins |
| Morse Max | $49 a month, or $468 a year | Everything in Pro, 5,500 monthly credits, encrypted sync on 5 devices, and 15 scheduled tasks |
Mammouth's quota page makes a fair point for its higher tiers, that "the Expert plan offers 10 times more quota for only 6 times the Starter price." Morse's prices are in US dollars, and its plans page lists what each plan includes.
How do Mammouth's usage limits work?
By session, and the session is 3 hours. Mammouth's quota policy says "Usage limitations on Mammouth are defined per session. Quotas are fully renewed after 3 hours." What counts against the quota is more than the number of messages: message and reply length, attached files, the model, tools like web search, Mammouth's own instructions, and conversation length, since "The longer a conversation gets, the more costly each exchange becomes."
When you reach a model's limit, Mammouth keeps you going on a smaller model. Its policy calls this the model threshold principle: "Mammouth automatically switches to a lighter model for the rest of your exchanges," and its example runs from Claude Opus to Claude Sonnet to Claude Haiku. The pricing page describes the same thing from the other side: "If a model hits its limit, lighter models stay available longer." Mammouth doesn't publish how many messages a quota holds, only that Standard has 3 times Starter's and Expert 10 times.
Morse Pro counts differently: unlimited messages on Pro models such as Claude Sonnet 5.5, Grok 4.7, and DeepSeek V4.1 Flash Uncensored, plus 250 monthly credits for the top models such as Claude Fable 5.1 and GPT-6 Astra. Max adds 5,500 monthly credits.
Which has more models, Mammouth or Morse?
Morse has more by count, with 80+ models, open and closed, in one account. Mammouth's home page Q&A lists 21 text and search models, among them GPT-6 Astra, Claude Opus 5.5, Claude Fable 5.1, Gemini 4 Argon, Grok 4.7, Mistral Large 4, DeepSeek V4.1 Flash, Kimi K3.1 Pro, GLM 5.3 Flash, and Perplexity Sonar Pro, plus 4 image models and 3 video models. Mammouth's own app sends one prompt to another model in a click, which it calls "One-click reprompting." Mistral's own app is reviewed in is Le Chat private.
On Pro and Max, Morse adds uncensored models, for answers other apps refuse: dark fiction, straight medical and legal answers, and help testing your own security.
How Morse keeps a model subscription from holding your chats
Each Mammouth issue above traces to one design choice. Your conversations live on Mammouth's servers for as long as you subscribe, while each prompt goes to a provider that may keep it for up to 30 days. Morse is built so your history is stored in your browser, not on its servers, and you pick per chat who gets to read the prompt.
- Anonymous, for Claude, GPT, and Gemini without your account behind them. Morse sends your message under its own account, so the model provider never learns your name, email, account, or IP address. The provider reads the message without knowing it's you, and may keep it for a limited time to check for abuse, under its own policy, so leave names and identifying details out.
- Private, the default, for models whose hosts agreed to keep nothing. Open-weight models like DeepSeek, Kimi, and GLM, plus xAI's Grok, run here. Morse has zero data retention agreements with its inference, search, and other providers, so they've agreed to keep nothing, and where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't. For web search, use a Private chat.
- E2EE, for prompts you want sealed before they leave your browser. Your browser checks the enclave's hardware attestation, then encrypts the message so only that verified enclave can read it. Morse's servers relay text they can't read. Every E2EE reply has a lock you can open to see each check that passed. E2EE chats come with Pro.
Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse's servers keep no log of your messages. Morse doesn't train AI on your chats, sell your data, or show you ads. Because Morse doesn't store chats, it usually has very little to give authorities: what it could be required to hand over is limited to account records. Your device holds your history, so save a backup to a file from Settings > Data. The privacy page shows what each party sees in each mode.
Mammouth or Morse: which should you choose?
- Choose Mammouth if you want Claude, GPT, Gemini, Mistral, and Grok on one bill from $10 a month paid yearly, a French company under the GDPR with your history hosted in the EU, and one-click reprompting to try the same question on another model, and you're fine with your conversations on Mammouth's servers and your prompts kept by providers for up to 30 days.
- Choose Morse if you want 80+ models where you decide per chat who reads the prompt, a chat history stored in your browser instead of on a company's server, E2EE that your own browser verifies, and uncensored models on Pro.
Send a first message from the Morse home page without an account, or start free with no card. Other model hubs are compared in Poe alternatives and ChatLLM alternatives.
Sources
Read November 20, 2026.
- Mammouth: home page, pricing, privacy policy, terms of service, data processing addendum, acceptable use policy, and sub-processor list, Privacy with model providers, quota policy, FAQ
Morse publishes this comparison. We've tried to keep it fair; if something's wrong or out of date, email hello@trymorse.ai.
Mammouth is a trademark of Mammouth AI SAS. Morse is not affiliated with or endorsed by Mammouth AI SAS.
Common questions
Who makes Mammouth AI?
Mammouth AI SAS, a French company with its registered office in Paris. Its terms put contracts under French law and the courts of Paris. Morse is made by Watermark Labs, Inc., of San Francisco, and the two companies aren't affiliated.
Does Mammouth AI have a free plan?
Its pricing page lists none. The cheapest plan, Starter, is $12 a month before VAT, or $10 a month when you pay $120 a year in advance, and its terms mention free trials only 'under specific circumstances.' Morse has a free plan with no card needed.
Is Mammouth AI GDPR compliant?
Mammouth says it processes data under the GDPR and keeps personal data in the European Economic Area, using Standard Contractual Clauses when data leaves it. Its own sub-processor list places several of its AI model providers in the USA.
Can I delete my Mammouth chat history?
Yes, from your account settings. Its privacy policy says deleted data may be kept for up to 30 rolling days for abuse detection, and account data for up to a year after your subscription ends.
What happens when you reach Mammouth's usage limit?
Mammouth switches you to a lighter model for the rest of the session, for example from Claude Opus to Claude Sonnet, then to Claude Haiku. Quotas renew in full after 3 hours.
Does Mammouth sell your data or show ads?
Its privacy policy says Mammouth makes money only from subscriptions and doesn't sell or share your data for advertising. Morse doesn't train AI on your chats, sell your data, or show you ads.
Is Mammouth cheaper than Morse?
Yes, at the entry level. Mammouth Starter is $12 a month before VAT, against $19 a month for Morse Pro, though Morse also has a free plan and Pro includes unlimited messages on Pro models.
Where does Morse keep my chats?
Your chat history is stored in your browser, encrypted, not on Morse's servers, and Morse's servers keep no log of your messages. Save a backup to a file from Settings > Data so a lost device doesn't take your chats with it.