AI Features Terms
What these features are
| Feature | What it does |
|---|---|
| Auto-Prompt | You write an instruction now. At a scheduled future time, an AI model runs it and delivers the result to a chat, a Channel, or verified email addresses. |
| Suggested messages | Pre-written and AI-assisted templates offered while you compose a Tribute. |
| Horaro AI | Our collective name for the above and any AI features we add. |
ℹ️ Memorial Video is provided by a separate application, launched from Let's Horaro and governed by its own terms and privacy policy. These Terms do not cover it.
The thing that makes this different
Most AI tools give you output immediately, and you decide whether to use it.
Auto-Prompt does not work that way. You write an instruction today; it runs on a date you choose — possibly months or years from now — and the result goes straight to its destination.
That means:
- You will not see the output before it is delivered.
- We do not review it before it is delivered.
- The world may have changed between writing and execution. Facts go stale. Circumstances change. A prompt that made sense in 2026 may produce something inappropriate in 2031.
- The underlying AI model will probably have changed, and may behave differently.
Please think about this before scheduling an Auto-Prompt to something that matters. A scheduled AI message to a whole company, or to a bereaved family, carries real risk of landing badly.
Our advice: test your prompt first with a near-term execution to yourself, keep prompts specific and self-contained, avoid prompts depending on facts that will age, and don't schedule AI output to sensitive audiences without a human in the loop.
What we don't promise about AI output
AI models generate plausible text and media. They do not verify truth.
Output may be:
- factually wrong, including confidently wrong
- outdated
- biased or culturally inappropriate
- offensive, despite our safety filters
- nonsensical or off-topic
- different every time, even from the same prompt
We provide AI features "as is". We do not warrant that output will be accurate, appropriate, complete, or fit for any purpose.
Do not rely on AI output for
- ❌ Medical, health or mental health decisions
- ❌ Legal advice or documents
- ❌ Financial, tax or investment decisions
- ❌ Safety-critical decisions
- ❌ Anything where being wrong causes real harm
If you are dealing with grief, distress or a mental health difficulty, please speak to a qualified person, not an AI feature.
Human review is your job
Where AI output goes to an audience — a Channel, a company group, a distribution list — you are responsible for it as if you had written it yourself. Our Acceptable Use & Community Guidelines apply to AI-generated content exactly as they apply to content you type.
How we handle your data in AI features
What gets sent to AI providers
Only what the feature needs:
| Feature | Sent |
|---|---|
| Auto-Prompt | Your prompt text and any context you include |
| Suggested messages | The occasion and relationship type — not your draft text unless you ask for a rewrite |
We do not send AI providers: your contacts, your chat history, your end-to-end encrypted messages, your payment details, or any Tribute content other than what you submit to the feature.
Where processing happens
Some AI providers process outside the UAE. Where they do, we rely on the transfer safeguards in Section 9 of our Privacy Policy. Current providers and their locations are listed at letshoraro.com/legal/sub-processors.
If you would prefer nothing left the UAE, don't use AI features. The rest of the Platform works without them.
Training
We do not use your content to train AI models, and we contractually require our AI providers not to train on data we send them.
Retention
AI prompts and outputs are retained under the retention schedule in our Privacy Policy. Prompts are retained until the Auto-Prompt is deleted or executed for the last time. A scheduled Auto-Prompt necessarily means we hold your prompt text until it runs — that is unavoidable given what the feature does.
Who owns AI output
You own the output, to the extent it is capable of being owned, and subject to two things you should know:
- AI output may not attract copyright. In many jurisdictions, material generated without sufficient human authorship is not protected. We cannot promise you own it in the way you own something you wrote.
- Output is not unique. Another user submitting a similar prompt may receive similar output. We cannot promise exclusivity.
You are responsible for checking that output does not infringe anyone else's rights before you publish or rely on it.
Auto-Prompt email destinations
You can send Auto-Prompt output to email addresses.
- New addresses are added as unverified and receive nothing until verified.
- Verification is a confirmation link sent to that address.
- Every Auto-Prompt email contains an unsubscribe link. Clicking it removes that address from all Auto-Prompt destinations on your account, immediately and permanently.
- You must have a proper basis for sending to an address that is not your own. Adding colleagues, clients or contacts without their agreement is spam and breaches our Acceptable Use & Community Guidelines.
An unsubscribe cannot be overridden by re-adding the address. It stays suppressed until the owner of that address asks us to lift it.
Credits and failures
AI features are charged in Credits. The rates are shown in the app before you confirm.
If an execution fails because of a technical failure on our side, we refund the Credits for that execution and retry once.We do not refund Credits because you didn't like the output. Output quality is inherently variable, and dissatisfaction with a result is not a service failure.
If an Auto-Prompt fails repeatedly, we will disable it and notify you rather than consuming your Credits on a loop.
Prohibited uses
Do not use AI features to:
- generate content that breaches our Acceptable Use & Community Guidelines
- generate content unlawful in the UAE
- impersonate a real person, or generate synthetic likenesses without authority
- generate misinformation for distribution
- generate spam or bulk unsolicited messaging
- attempt to extract training data, system prompts, or model internals
- circumvent safety filters, including by obfuscating a prohibited request
- attack, reverse engineer or benchmark our AI providers' systems in breach of their terms
- generate anything involving the sexualisation of minors — zero tolerance, immediate termination, reported
Attempts to circumvent safety filters may result in loss of access to AI features, and repeated attempts may result in account termination.
Changes to AI features
AI models and providers change frequently. We may change, replace or withdraw the model behind a feature without notice, which may change the character of the output.
If we withdraw an AI feature entirely, we will give 30 days' notice, refund Credits you bought specifically for it, and notify you of scheduled Auto-Prompts that will no longer execute so you can make other arrangements.
Liability
Section 17 of our Terms of Service applies. In particular:
- We are not liable for the content of AI output.
- We are not liable for decisions you or anyone else makes based on AI output.
- We are not liable for consequences of an Auto-Prompt delivering content you would not have sent had you seen it first.
This is not us disclaiming everything and walking away. We will fix failures, refund Credits for executions that break, and tell you honestly when something has gone wrong. But the judgement about whether to schedule an unreviewed AI message to a real audience is yours, and it has to be.
Questions
support@letshoraro.com