Sumchat AI Services Terms

Effective date: September 27, 2026 · Version 1.0

Sumchat is built around AI: Agents in your workspace draft, build, edit, publish, and communicate at your direction. These AI Services Terms explain how that works legally; who owns what, what we promise and don't, and the rules for AI-made content. They are part of the Terms of Service.

1. Definitions

"Input" means the prompts, briefs, conversations, files, and other material you and your members provide to Agents. "Output" means content an Agent or other AI feature of the Services generates in response; text, code, sites, images, audio, video, plans, and documents.

2. Ownership of Output

2.1 Yours, as between us. As between you and Sumchat, you own all right, title, and interest, if any, in your Input and Output. To the extent Sumchat holds any intellectual-property rights in Output, Sumchat assigns them to you when the Output is generated. Output you use, save, approve, or publish is Customer Content under the Terms.

2.2 What "if any" means. United States law requires human authorship for copyright. Purely machine-generated material may not be protectable by anyone, and a prompt alone does not make you its author. Your own creative contributions; your direction, selection, arrangement, editing, and the material you supply; are what make a finished work yours to protect. We make no promise that any Output is protectable, registrable, original, or unique: similar requests from other customers may produce similar output, and only your copy is assigned to you.

2.3 No reuse across customers. We do not take Output generated for your workspace and provide it to other customers as theirs, and we do not use your Input or Output to train models (Terms § 11.2). Templates, scaffolding, and platform components we provide to everyone remain platform materials (Terms § 4.4).

3. Your responsibilities for Output

3.1 Review before you rely or publish. AI output can be inaccurate, incomplete, out of date, biased, or similar to existing works, and it can state things confidently that are wrong. You are responsible for reviewing Output before relying on it or publishing it. Approval gates exist so that a person in your workspace confirms consequential actions; what you approve and publish is yours.

3.2 Not professional advice. Output is not legal, medical, financial, tax, or other professional advice, and no professional relationship is created by using the Services. Sumchat is not a law firm and does not practice law. Where the Services generate documents of a legal nature; privacy policies, terms pages, disclosures; they are general self-help templates populated with information you provide, not documents prepared for your particular legal situation; they are not reviewed by an attorney, we express no opinion on their legal sufficiency, and you should have your own attorney review them. You control whether to use them.

3.3 Compliance reviews are information, not clearance. Where the Services review a site or content for potential compliance issues (for example, accessibility or required-disclosure checks), the review is informational tooling: it can find some problems, it cannot certify the absence of problems, and it is not a legal determination that anything complies or fails to comply with any law. Passing a review transfers none of your Terms § 8.2 obligations to Sumchat.

3.4 Rules for AI content. You may not use AI features to violate the AUP, and in particular you may not: generate fake reviews, testimonials, or endorsements; impersonate real people or organizations, including through voice or likeness; present AI to consumers as human; generate deceptive or fraudulent material; or produce political advertising through the Services' publishing features. You must not remove, strip, or defeat provenance metadata, watermarks, or AI-content labels that the Services or upstream providers embed in generated media, and where a social network requires AI-generated content to be labeled, you must not defeat the label.

4. How the Services use AI providers

Agents run on third-party foundation models and AI infrastructure (see the Subprocessors list, including OpenAI for the assistant, hosted computer work and voice, Google Gemini for media work, and TypeSafe for interpreting requests). Your Input and Output are routed to these providers to provide the Services, under agreements that prohibit them from training on your content. Provider availability and behavior are outside our control: models are updated by their developers, may be refused or filtered by provider safety systems, and may occasionally be unavailable. A safety refusal or model change is not a defect in the Services.

5. Voice and calls

AI voice features identify themselves as AI and disclose recording at the start of calls they answer or place. You must not instruct an Agent to conceal that it is an AI, and outbound calling is subject to the consent obligations in Terms § 7. Call audio and transcripts are processed only to provide the Services to your workspace; never for training, and never shared across tenants. You may not use voice features to clone or imitate the voice of any real person.

6. Memory

Agents retain conversation context and saved memory to work effectively across conversations. This is Customer Content, scoped to the relevant personal or workspace context and excluded from model training. You can request access, export, correction or deletion through the contact in the Privacy Policy. Google disconnection alone does not erase content already saved in this context. Details are in the Privacy Policy.

7. Indemnification note

Your indemnification obligations under Terms § 15 apply to claims arising from Output you use or publish. Sumchat does not indemnify you for third-party claims that Output infringes; Section 2.2 explains why we cannot promise Output is non-infringing, and Section 3.1 is the control that protects you.

8. Changes

AI capability evolves quickly; we may update these AI Services Terms per the Terms' change-notice process.

Questions: legal@sumchat.ai