Terms of Service
Effective Date: [NEEDS DECISION: set before the alpha is exposed to any user not personally known to the team — align with the Privacy Policy effective date] Last Updated: October 1, 2026 (V0.3.1 — licensing reference aligned)
Status: DRAFT — not yet effective. Rewritten to match the service FactHarbor actually operates today: an invite-gated, AI-assisted claim-analysis service in alpha, not a community-contribution platform with user accounts. Items marked [NEEDS DECISION: …] require a human/legal decision before adoption and entry into effect; this draft may be published for review. Reviewed internally and by a GPT-5.5 counsel-style pass; not yet signed off by retained Swiss counsel (a launch blocker — see public launch requirements).
1. Acceptance of Terms
These Terms of Service and our Privacy Policy govern your use of FactHarbor. To ensure these Terms are validly incorporated, we present them — together with the third-party-processing and visibility notices (Privacy Policy §§3.3, 5–7) — for your acceptance before you submit a claim; by submitting, you confirm you accept them. If you do not agree, do not use the service. [NEEDS DECISION: implement a click-through/checkbox acceptance ("Submit claim and accept Terms") — mere "by using" (browsewrap) is weak for incorporating the liability, indemnity, public-visibility, and third-party-transfer clauses, and weaker still for consumers. For current known testers, obtain acceptance in writing via the invitation flow.]
2. Description of the Service
FactHarbor is operated by a Swiss non-profit association (Verein), UID CHE-448.446.098, Canton of Zürich.
Current (alpha) functionality: FactHarbor is an AI-assisted claim-analysis service. Using an invite code, you submit a factual claim (as text or a URL); our automated pipeline researches it using web search and large-language-model (LLM) reasoning and returns a structured analysis report.
Important — how the analysis works: to produce a report, FactHarbor sends content derived from your submission to third-party AI and search providers and fetches third-party web pages. See Privacy Policy §5 for the providers involved and §6 for the sensitive-data warning. Do not submit content you are unwilling to have processed this way.
The account-based community features (public contributions, profiles, collaborative editing) described in earlier drafts are planned for a later phase and are not active. These Terms will be updated before those features launch.
3. Alpha / Experimental Status
The service is provided on a pre-release, experimental basis for evaluation:
- Features may change, break, or be withdrawn at any time without notice.
- We do not guarantee availability, performance, or that data (including your submissions and generated reports) will be retained or preserved.
- There is no service-level agreement and no commitment of support response times during alpha.
4. Access and Invite Codes
- Access requires a valid invite code that we issue. Invite codes are subject to usage quotas.
- Invite codes may be used by the people we issue them to; you may not publish, sell, or pass them to unauthorised persons, or attempt to circumvent quotas, rate limits, or other access controls.
- We may revoke or disable any invite code, or suspend or terminate access, at any time — including for violation of these Terms, abuse, security risk, or where required by law.
- The alpha does not use accounts; you do not register credentials, and we do not authenticate you as an identified individual.
5. Acceptable Use
You agree to:
- Use the service in good faith and for its intended purpose (evaluating factual claims).
- Comply with all applicable laws.
You may not:
- Submit content that is unlawful, infringing, or that you do not have the right to submit.
- Submit personal or sensitive data about yourself or others except as essential to the claim, understanding it will be processed as described in the Privacy Policy (§5–§6). If you submit a claim about an identifiable person, you are responsible for having a lawful basis to do so.
- Submit claims that unlawfully infringe a third party's personality rights (ZGB Art. 28) or are defamatory, or use the service to harass, defame, or harm others, or to generate or launder disinformation.
- Attempt to manipulate, overload, reverse-engineer, or abuse the analysis pipeline, or circumvent access/rate controls.
- Use automated tools to submit at scale beyond your invite quota.
6. Content, Submissions, and Intellectual Property
6.1 Your Submissions
- You retain ownership of the content you submit.
- You represent that you have the right to submit it and to permit its processing as described.
- You grant FactHarbor a limited, non-exclusive licence to use, reproduce, transmit, and process your submission — including transmitting it to the third-party AI/search providers and source-fetching described in the Privacy Policy — solely to provide and improve the analysis service. We do not permit providers to use your submissions to train their models except where a provider's configured terms allow it. [NEEDS DECISION: confirm provider training/retention settings per provider; if "improve" ever includes our own training/evaluation datasets, say so explicitly here.]
6.2 Analysis Reports
- Analysis reports are generated by FactHarbor's automated pipeline. See the disclaimer in §7.
- Reports may concern identifiable persons or organisations; we apply the personality-rights safeguards and complaint/takedown route in Privacy Policy §6.2.
- FactHarbor deliberately makes submissions and generated reports public through job-list and job-view interfaces reachable without an invite code, including the full submitted text or URL and the full generated report. Do not submit confidential, private or unnecessary personal information, or anything you are unwilling to make public. See Privacy Policy §3.3.
- When you copy or republish a report, you must preserve its disclaimers, not edit it misleadingly, and not use it to harass, defame, or mislead.
6.3 Platform and Trademarks
- Public source code is available under AGPL-3.0-only and public documentation under CC-BY-SA-4.0, subject to the path-specific coverage and notices in Open Source Model and Licensing.
- The FactHarbor name and marks remain our property.
7. AI-Generated Output — No Guarantee of Accuracy
This is critical to understand:
- Analysis reports are produced by automated AI reasoning over automatically retrieved sources. They may contain errors, omissions, outdated information, or fabricated/“hallucinated” statements, and may misattribute or misinterpret sources.
- A report is an automated assessment, not a definitive determination of truth, and does not constitute professional, legal, medical, financial, or any other expert advice.
- You must exercise your own judgment and independently verify anything you rely on. Do not make consequential decisions based solely on a FactHarbor report.
- Verdicts, truth percentages, and confidence figures are machine-generated indicators, not authoritative rulings on any person, organisation, or statement.
8. Disclaimers
8.1 "As Is"
The service is provided "as is" and "as available", without warranties of any kind, express or implied, including merchantability, fitness for a particular purpose, accuracy, and non-infringement.
8.2 Content Accuracy
We do not guarantee the accuracy, completeness, or reliability of any analysis report or of user-submitted content (see §7).
8.3 Availability
We do not guarantee uninterrupted or error-free operation (see §3).
9. Limitation of Liability
To the maximum extent permitted by Swiss law, FactHarbor and its officers, members, and operators shall not be liable for any indirect, incidental, special, consequential, or punitive damages, or for any loss of data, arising from your use of (or inability to use) the service or from reliance on any analysis report. Nothing in these Terms excludes or limits liability for intent or gross negligence (OR Art. 100), for death or personal injury, or any other liability that cannot be excluded under mandatory law. For consumers, the limitations apply only to the extent permitted by mandatory consumer-protection law.
10. Indemnity
You agree to indemnify and hold FactHarbor harmless from third-party claims arising out of content you submit in culpable breach of §5 or §6.1 (e.g., unlawful, infringing, defamatory, or unauthorised personal data). For consumers, this applies only to the extent permitted by mandatory law. [NEEDS DECISION: confirm whether to retain this clause for an alpha with known testers, or defer to launch.]
11. Age and Capacity
You must have the capacity to enter these Terms under the law applicable to you. The service is not intended for children: under Swiss law a minor may agree only if capable of judgment (Urteilsfähigkeit, ZGB Art. 16); where the GDPR applies and consent is the basis, the age is 16 (Art. 8; a member state may lower it, but not below 13); US COPPA uses 13. See Privacy Policy §15. [NEEDS DECISION: consider requiring a minimum age of 16 or 18 for a public consumer launch for simplicity.]
12. Termination
You may stop using the service at any time. We may suspend or terminate your access (including invite codes) at any time as described in §4. Provisions that by their nature should survive termination (§§6–10, 13) survive.
13. Governing Law and Jurisdiction
These Terms are governed by Swiss law. The place of jurisdiction for disputes is the Canton of Zürich, Switzerland, subject to any mandatory consumer-protection jurisdiction and law that applies to EU/EEA (or other) consumers. If any provision is held invalid, the remaining provisions continue in full force.
14. Changes to These Terms
We may update these Terms. Material changes will be announced with reasonable notice via a prominent notice on the site (and by email where we hold one). Continued use after a change takes effect constitutes acceptance. Previous versions will be archived.
15. Contact
For questions about these Terms: FactHarbor (Verein), c/o Robert Schaub, In Lederäcker 11, 8305 Dietlikon, Switzerland; info@factharbor.ch.
16. Version
- Version: 0.3.1 (licensing reference aligned), superseding the unnumbered December 17, 2025 draft.
- Last Updated: October 1, 2026
- Effective Date: [NEEDS DECISION — see header]
Changelog:
-
0.3.1 (2026-10-01): updated §6.3 to the adopted repository licensing and removed licensing from the open-decision list. This remains a draft without an effective date.
-
0.3.0 (2026-09-17): recorded full submissions and reports as deliberately public by design; replaced the visibility and report-reuse decisions with a direct confidentiality warning and binding reuse conditions; clarified that the draft may be published for review before adoption; and added the serviceable contact address.
- 0.2.1 → 0.2.2 (GPT-5.5 counsel review): added a click-through-incorporation requirement (§1); preserved mandatory liability under OR Art. 100 + consumer caveat (§9); narrowed the indemnity to culpable breach + consumer caveat (§10); added a personality-rights/no-defamation duty (§5) and a cross-reference to the Privacy Policy §6.2 takedown route (§6.2); added report-use rules for shared reports (§6.2); made the provider-training position explicit (§6.1); flagged a possible 16/18 minimum age for public launch (§11).
- 0.2.0 → 0.2.1 (Fable-5 advisor review): children's-age clause fixed; invite-code wording aligned to Privacy Policy §3.1; §6.3 "will be released"; visibility pointer to Privacy Policy §3.3.
Open decisions before adoption and use as effective terms: §1 click-through acceptance; §6.1 provider training/retention; §10 indemnity scope; §11 minimum age; effective date. The broader compliance checklist lives in the Privacy Policy §20.