Effective date: [YYYY-MM-DD] · Version: [YYYY-MM-DD]
This policy is part of the SplitSpoon Terms of Service (the "Terms of Service"). Words in quotation marks have the meaning they have in the Terms of Service. [SplitSpoon LLC], a [State] limited liability company, is "we" or "us". If this policy and the Terms of Service conflict, the Terms of Service control; the order of precedence is in Terms of Service section 17. [lawyer: precedence; see Terms of Service section 17]
1. Your content
- Upload only documents you have the right to submit.
- If a document belongs to or was made for your client, you must have your client's permission to upload it. Any confidentiality terms your client has set are yours to honour. We do not see those terms and cannot check them.
- Do not upload anything unlawful, or anything that infringes someone else's rights.
- Do not upload malware, or files made to harm the Service or other people.
- The Service is for subsurface documents such as boring logs. Do not use it for anything else that breaks this policy.
2. Accounts
- One account is for one person. Do not share a sign-in, a password, a vault passphrase or a recovery key.
- An invitation code is personal to the person we invited. Do not pass it on, post it, or sell it. A code expires as Terms of Service section 2 says. Invitations normally bind the code to an email address. A bound code works only with that address and is mailed there; an operator can explicitly issue an unbound invitation. [verify: bound mail and the safer CLI default are built on
prep/int, not yet deployed.] - Do not resell access to the Service, or let someone else use your credits. Credits cannot be moved to another account.
- Do not hold more than one account unless we agree.
- Do not create accounts to get around a limit, a suspension or a deletion.
- Business use only. You must be 18 or older.
3. Security
Do not:
- probe, scan or test the Service for weaknesses without our written permission, other than good-faith research as described below;
- overload the Service, or send traffic meant to slow it down or stop it;
- get around encryption, authentication, rate limits, spending limits, quotes or account restrictions;
- try to reach another customer's account, projects or data;
- reverse engineer the Service. [lawyer: scope of the reverse-engineering ban; see Terms of Service section 8]
Good-faith security research. If you find a weakness, tell us at [email protected]. We acknowledge reports within two business days, fix confirmed issues before disclosing them, and credit reporters who want it. Please:
- report it to us first, and give us time to fix it before you say anything publicly;
- test only against your own account and your own data;
- stop as soon as you reach another customer's data, and tell us;
- do not degrade the Service or keep any data you should not have seen.
The safe harbour covers testing to find and report a weakness. It does not cover testing for competitive analysis. We will not treat research that follows these rules as a breach of this policy. [lawyer: safe-harbour wording and whether to promise not to bring claims] [owner: confirm there is no bounty]
4. Fair use and limits
The Service has limits that protect the Service and your balance. These are the values the code uses today. We may change them, and we may raise or lower them for one account. Several are settings, so they may differ for your account or at a later date. [verify at launch]
| Limit | Value today |
|---|---|
| Pages in one run | 60 (per-account setting) |
| Pages per rolling 24 hours | 300 (per-account setting) |
| Runs at the same time | 2 (per-account setting) |
| Files in one request | 50 |
| Stored project data per account | 5 GB (per-account setting) |
| One uploaded document | 50 MB by default (a server setting) |
| Total uploaded data per project | 64 MiB at most (a server setting, clamped) |
| Size estimates running at once | 2 workers, with a queue of 8 waiting (a server setting). Past that the Service answers "busy" at once. |
| Spending in the first 7 days of a new account | 10,000 credits (US$100) by default; we can raise it on request |
| Other requests | Request size caps, and per-account rate limits on sign-in-protected routes such as address lookups, upload signing, status polling, project saves, vault writes and checkout. Sizes and rates are server settings. |
| Public invitation-request form | 5 requests per 10 seconds per source, at our network edge |
If you hit a limit, wait and try again, or write to [email protected] if you need more. Do not work around a limit by opening more accounts or by sending requests in a way meant to avoid it.
Reasonable use for your own work is fine. We may restrict use that is far beyond the normal pattern for a customer, or that puts the Service at risk for others. [lawyer: "fair use" wording for paid, prepaid credits]
5. AI-specific rules
- Do not try to make the AI reveal another customer's data, or its own instructions, prompts or settings.
- Do not use prompt injection: do not put text in a document meant to make the AI ignore its instructions, change its output for other people, or act beyond reading the pages you submitted.
- Do not use the Service or its output to train, tune or build a competing AI model, dataset or service. [lawyer: your own use of exports for your own work is allowed; check that this carve-out is clear]
- Do not use the Service to process documents for others as a paid extraction service without our written agreement. [owner: allow or forbid paid extraction bureaus; this is the proposed text]
- Do not scrape the Service, or use bots or scripts to submit or collect content in bulk, except through an interface we give you for that purpose. We do not currently offer an interface for scripted use. The Service's programming interface is for SplitSpoon's own web app and is not one. [owner: confirm; the code has no customer API keys or published API documentation, only the web app's own sign-in]
- Our AI providers (Anthropic and OpenAI) have their own usage policies. Do not use the Service in a way that breaks them. We may pass their rules on to you. [verify the providers' current policies and name the links on the publish date]
- Review AI output before you rely on it. See Terms of Service section 5.
6. Enforcement
If we reasonably believe you broke this policy, we may, depending on how serious it is:
- warn you;
- remove or block content;
- limit what your account can do;
- suspend or end your access (Terms of Service sections 9 and 9A), and delete your account after the notice Terms of Service section 9 requires;
- tell the authorities, and work with them, where we believe the use is unlawful or the law requires it. [lawyer]
We will usually show a reason in the app when we suspend an account. We may act without notice where there is a security risk or where we must. Suspension or deletion does not refund unused credits, except as the Refund & Credits Policy and the law require. [lawyer: consistency with the Refund & Credits Policy]
To report abuse or a breach of this policy, write to [email protected]. Security problems go to [email protected].
7. Changes
We may update this policy as the Terms of Service allow (Terms of Service section 15).