Morse vs Chai: who keeps your character chats, what trains on them, and what you pay
The Morse teamPublished Facts checked 9 min read
Morse vs Chai comes down to who holds your character chats. As of November 2026, Chai Research stores your conversations in its systems in the United States, fine-tunes its models on them and on the replies you prefer, and may keep your data for five years after you delete your account, while Morse stores your character chats encrypted in your browser, not on Morse's servers, and doesn't train AI on them. Choose Chai if you want its library of 25 million user-made characters, images and short videos of them in chat, and an Android app; choose Morse if you want unlimited character messages and swipes on a free plan, no ads, and character chats that stay encrypted in your browser.
We read Chai's privacy policy (last updated November 5, 2026), its end user license agreement (July 28, 2026), its safety center, company facts, and home page, plus the FAQ, pricing, and terms of service of Chai's web product and its Google Play listing, all on November 18, 2026. Every Chai fact below links to one of those pages. Chai deserves credit for a public safety center, and for web terms that let you cancel "at any time, online, without calling or contacting us" (Chai).
Morse vs Chai at a glance
As of November 18, 2026.
| Morse | Chai | |
|---|---|---|
| Where your prompt is decrypted | A character's reply comes from the model playing it; in Chat, you pick per chat: E2EE inside a verified enclave, Private at a host that agreed to keep nothing, or Anonymous at the provider, without knowing it's you | On Chai's systems, "processed and stored in the United States" (Chai) |
| Who verifies it | Your browser, in E2EE chats | Not described |
| Where chat history lives | Stored in your browser, encrypted, not on Morse's servers | "Stored in our systems," for up to five years after you delete your account, longer for legal claims (Chai) |
| Trains on your chats | Morse doesn't train AI on your chats | Yes, on conversations and "the preferences you express between responses" (Chai) |
| Ads | None, on any plan | For free users who consent; none on Web Pro (Chai) |
| Models | 80+, open and closed; on Pro, your choice of model | Chai's "in-house developed" models (Chai) |
| Free plan | Unlimited character messages and swipes, no card | "Free to start"; a paid plan is required in some regions (Chai) |
| Paid plans | Pro $19 a month ($17 yearly); Max $49 ($44 yearly) | Web Pro $9.99 a week or $159.99 a year; MAX add-on $90 a month or $0.66 per 1,000 tokens (Chai) |
| Age rule | Users confirm they're 18 or older | 18+, with selfie or ID checks (Chai) |
What changed in Chai's privacy policy on November 5, 2026?
Chai rewrote it to cover its web product. The October 12 version, in the Internet Archive's copy of October 26, spoke of "the Chai mobile applications" only; the November 5 version covers "the Chai web application and the Chai mobile applications for iOS and Android," and says "This policy applies to the web and mobile versions of the APP" (Chai), though on November 18, 2026, Apple's lookup still returned no listing for Chai's iPhone app in the US or the UK (Apple, Apple). Three other changes, none of them in the October 12 text, affect character chats.
- Training now names your picks. Chai fine-tunes its models, "including our safety models," on "your in-app conversations with chatbots and the preferences you express between responses." It removes personal identifiers first and stores that data "separately from your account information."
- The supplier list has names for its categories. Your data is shared with suppliers acting as data processors, and the policy now lists them as "hosting, AI model, media generation, age verification, analytics, notification, support and payment providers."
- Age checks have new routes. Users "may be asked to complete an age check by identity document"; on Android, Chai uses Google Play Age Signals, and on iOS it is "in the process of implementing the relevant Apple ID age information."
The five-year retention is unchanged: data is kept no longer than "the duration of the account and five (5) years following its deletion," and longer when Chai needs it for legal claims or orders (Chai).
Who does Chai's November policy share your data with?
With the suppliers its November 5 policy lists above, though its web FAQ still says "Chai AI does not share your personal data with any third parties" (Chai). How the FAQ and the policy disagree is in Chai AI alternatives. Morse doesn't sell data, runs no advertising trackers or third-party analytics, and has zero data retention agreements with its inference, search, and other providers, so in Private mode they've agreed to keep nothing. In Anonymous mode, the provider may keep a message for a limited time to check for abuse, under its own policy.
What does Chai learn when you swipe for a new reply?
Chai's policy doesn't use the word "swipe," but it trains on "the preferences you express between responses." Chai's home page traces its models to signals like that: a reward model trained "directly on user signals" in March 2023, a larger one "trained with 100 million signals to decrease user retry rate," and a 2024 model trained "with user preference datasets" that "boosted engagement by 20%" (Chai). Its company facts page says new models are "tested in A/B experiments with CHAI's more than 1M daily users" (Chai). Here is what each everyday action leaves behind:
| What you do | On Chai | On Morse |
|---|---|---|
| Send a message | Stored in Chai's systems and used to fine-tune its models, identifiers removed (Chai) | Stored in your browser, encrypted; Morse doesn't train AI on it |
| Pick one reply over another | A preference Chai can train on (Chai) | Unlimited swipes on Free; no training |
| Make a bot | Once posted, a "perpetual" license to Chai over your Contributions (Chai) | Characters you make are stored in your browser, encrypted |
| Generate an image or video | Counts as a Contribution under the same license (Chai) | Studio images are stored in your browser, encrypted |
| Delete your account | Data kept up to five years after deletion, longer for legal claims (Chai) | Chats are deleted, and Morse can't recover them |
In the EU or UK you can object to Chai's training, which rests on "legitimate interest"; Chai can refuse on "compelling legitimate grounds" (Chai). The policy describes no training switch. Chai's pages describe no chat export either, and the policy's right to data portability covers only people in the EU and UK (Chai). Which companies train on chats, and how to stop each, is in which AI companies train on your data.
Does Chai show ads?
On the free tier, with your consent (Chai), and not on Web Pro (Chai); is Chai AI safe covers the ad terms. Morse has no ads on any plan, Free included.
How old do you have to be to use Chai or Morse?
Eighteen, on both. Chai says it is "strictly for adults" and checks ages through Persona (Chai), as is Chai AI safe explains; Morse users confirm they're 18 or older, and sign in with a code sent to their email or with Google.
What safety features does Chai have?
Chai's November 5 policy adds two lines on enforcement: account suspensions are "reviewed by a person," and Chai reports to the National Center for Missing & Exploited Children (NCMEC) and authorities "where the law requires it" (Chai). Its self-harm scan, crisis notice, and content filters are in is Chai AI safe. In Morse's character chats, crisis lines appear when a message reads like suicidal thinking, and that check runs on your device. Morse reports CSAM to NCMEC.
How much does Chai cost compared with Morse?
It depends on how you pay Chai. Over a year, Chai Web Pro's annual plan, $159.99 or $13.33 a month (Chai), costs less than Morse Pro billed yearly at $17 a month ($204). Chai's weekly plan, $9.99 a week, comes to about $520 a year, more than Morse Pro at $19 a month ($228 a year). Full Chai pricing is in Chai AI alternatives.
Morse's Free plan costs nothing, needs no card, gives 15 messages a day with the model picked for you, and has no cap on character messages or swipes. Pro adds your choice of model and 3x the character memory; Max, at $49 a month, has 6x.
Which models does Chai use?
Its own. Chai's FAQ says it runs "in-house developed Large Language Models" on its own cluster of "over 5,000 GPUs" (Chai), and its company facts describe post-training on open models "including DeepSeek-V4 and Kimi-K2" (Chai). Its research on "blending" serves replies from several models "within one conversation" (Chai). Morse has 80+ open and closed models, GPT, Claude, Grok, DeepSeek, and Kimi among them, and Pro gives you your choice of model.
What does Chai do better than Morse?
Chai's strength is its library: "over 25 million characters created by users" and community rankings to find them (Chai). Pro subscribers can generate images and short videos of characters in chat, and an image package adds 50 a week (Chai). The web version also generates voice (Chai). Chai has an Android app on Google Play, too. If you're deciding between companion apps more broadly, our Chai AI alternatives guide compares five.
How Morse keeps your character chats in your browser
Every row above traces back to Chai holding your conversations on its own systems, which lets it decide what trains on them, who processes them, and how long a copy lasts after you leave. On Morse, your character chats, the characters you make, and your personas are stored in your browser, encrypted, not on Morse's servers. Our servers don't keep messages or log their content, and your account holds sign-in details, your plan, your balance, and a usage ledger with no content in it (privacy policy).
Morse doesn't train AI on your chats. Clearing your browser or deleting your account deletes your chats, and Morse can't recover them, so copy out any story worth keeping. Because Morse doesn't store chats, it usually has very little to give authorities. Moving from Character.AI too? Morse's importer brings your characters, chats, and personas over free on every plan.
In Chat, you pick who can read each conversation:
- E2EE, on Pro, for a regular chat Morse's servers should never be able to read. Your browser checks the enclave's hardware attestation, then seals the message so only that enclave can open it, and Morse's servers relay text they can't read. That rests on the enclave hardware and attestation working as designed, and E2EE chats have no web search or image generation.
- Private, the default, for open models such as DeepSeek, Kimi, and GLM. 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.
- Anonymous, for GPT, Claude, or Gemini without a login attached. Our proxy removes your account, email, and IP address, so the provider sees the message without knowing it's you, and may keep it for a limited time to check for abuse, under its own policy.
How the enclave check works is in end-to-end encrypted AI chat, and what Chai's own pages say about its 2023 case in Belgium and its old Google Play notice is in is Chai AI safe.
Morse vs Chai: which should you choose?
- Choose Chai if you want 25 million community characters, images and videos of them in chat, voice on the web, or an Android app, and you're fine with Chai keeping and training on your chats.
- Choose Morse if you want unlimited character messages and swipes without paying, no ads, character chats stored encrypted in your browser with no training on them, and, on Pro, your choice of model.
Start free and build your first character, with no card needed.
Morse publishes this comparison. We've tried to keep it fair; if something's wrong or out of date, email hello@trymorse.ai.
Chai is a trademark of Chai Research Corp. Morse is not affiliated with or endorsed by Chai Research Corp.
Common questions
Can I move my Chai characters to Morse?
By hand. Chai's pages describe no export tool, and its privacy policy offers data portability only to people in the EU and UK, so copy a character's description and greeting into a new Morse character. If you also have characters on Character.AI, Morse's importer brings those over free, with their chats and your personas.
Does Chai train on the reply I pick?
Chai's privacy policy, updated November 5, 2026, says it fine-tunes its models on your conversations and 'the preferences you express between responses,' with personal identifiers removed first. Morse doesn't train AI on your chats.
Is Chai or Morse cheaper?
It depends on how you pay. Chai Web Pro's annual plan is $159.99, or $13.33 a month, against Morse Pro's $17 a month billed yearly; Chai's weekly plan, $9.99 a week, costs more than Morse Pro's $19 a month. Morse's Free plan has no cap on character messages or swipes.
Are Chai and Morse for adults only?
Yes. Chai says it is strictly for adults and checks ages through Persona. Morse users confirm they're 18 or older.