Acceptable use
Acceptable Use Policy
Every question costs real money in model API fees, paid by one person. These rules exist so that the free tier can stay free and the paid tiers can stay cheap. Normal human use never bumps into them.
1. Technical limits
Limits apply per account and per IP address, whichever is stricter:
| Control | Default | Why |
|---|---|---|
| Burst limit | 6 asks per minute | Stops scripted floods |
| Hourly limit | 40 asks per hour | Caps sustained automated use |
| Minimum spacing | 3 seconds between asks | Distinguishes typing from scripting |
| Concurrency | 1 in-flight ask at a time | Prevents parallel fan-out |
| Monthly allowance | Per plan, shown in the app | Keeps costs predictable |
| Fair-use ceiling on “unlimited” | Internal | Backstop against runaway automation |
Hitting a limit returns a clear message telling you how long to wait. Limits may be tuned without notice as costs and traffic change. Repeatedly hammering a limit is treated as automated abuse.
2. Prohibited conduct
Automation and extraction
- No bots, scripts, scrapers, headless browsers, browser-automation extensions or scheduled jobs submitting queries.
- No attempt to enumerate, mirror, bulk-download or reconstruct the transcript corpus, the search indexes, or any substantial portion of them.
- No using output to train, fine-tune, distil, benchmark or evaluate a machine-learning model, or to assemble a competing dataset or product.
- No reverse-engineering the API for use outside the official interface, and no reselling access.
Circumvention
- No multiple or throwaway accounts to multiply free allowances, and no deleting and recreating an account to reset counters.
- No sharing an account, credential or session between people.
- No proxy, VPN rotation or IP cycling used specifically to evade limits.
- No tampering with quota, plan or entitlement responses.
Security
- No probing, scanning, load-testing, denial-of-service, or attempts to access data or systems you are not authorized to access.
- No injecting malicious payloads, and no attempts to make the model reveal server configuration, credentials or other users’ data.
- Good-faith vulnerability research is welcome — report privately to [email protected] before disclosing.
Content and conduct
- No unlawful use, and no using the Service to harass, threaten, stalk, defame or violate anyone’s rights.
- No publishing answers in a way that misrepresents them as official statements of America First, Nicholas J. Fuentes or the Operator.
- No stripping citations in order to pass output off as something other than what it is.
- No using share links to distribute unrelated spam or malware.
3. Enforcement
Depending on severity, we may:
- Throttle or temporarily block requests;
- Email you and ask you to stop;
- Suspend the account pending explanation;
- Terminate the account — without refund where the cause is a breach;
- Block an IP range or report unlawful activity to the appropriate authority.
Honest mistakes get a warning. If you think enforcement was wrong, email [email protected] and we will look at it properly.
4. Researchers, journalists and heavy users
If you have a legitimate need that these limits get in the way of — academic research, journalism, an accessibility requirement, or a project you would like to build on top of this — just ask. Email [email protected] describing what you need and why. Permission is often free; unauthorized scraping never is.