SKRIB, INC.

AI Terms (a Product-Specific Term)

Effective Date: 26 May 2026 | Last Updated: 26 May 2026

These AI Terms (the “AI Terms”) are a Product-Specific Term under the Skrib Terms of Service and the Skrib Software Services Agreement and govern your use of the AI features (the “AI Features”) of skrib.com and the related online services (the “Services”). Capitalised terms used and not defined here have the meanings given in the Terms or SSA.

In the event of conflict between these AI Terms and the Terms of Service or SSA with respect to AI Features, these AI Terms prevail.

1. Optional nature

AI Features are optional. You can use Skrib for the full set of writing, collaboration, and storage capabilities without ever invoking an AI Feature. Where AI is used, you invoke it; you decide what context is shared with the AI; and you review the output before relying on it. Account-level toggles to disable AI Features are available in account settings.

2. AI providers

Skrib uses third-party AI providers to deliver AI Features. The current providers and the scope of their use are summarised below. Skrib may add, replace, or remove providers from time to time as the Services evolve; the current list is maintained at this URL and reflected in the Subprocessor List at https://skrib.com/legal/subprocessors.

OpenAI, L.L.C. — Models / API: Enterprise / API (GPT-4 family); Used for: Drafting, editing, summarisation, brainstorming; Provider privacy policy: https://openai.com/policies/privacy-policy.

Anthropic, PBC — Models / API: Claude API (Claude Opus, Sonnet, Haiku); Used for: Drafting, editing, long-context analysis, summarisation; Provider privacy policy: https://www.anthropic.com/legal/privacy.

Google LLC — Models / API: Gemini API; Used for: Drafting, editing, multimodal assistance; Provider privacy policy: https://policies.google.com/privacy.

3. No training on user content

Skrib does not use user prompts, uploaded files, document content, or AI outputs to train Skrib’s own AI models.

Skrib uses enterprise or comparable API arrangements with each of OpenAI, Anthropic, and Google under which user content submitted through the Services is contractually prohibited from being used to train those providers’ foundation models. The relevant contractual restrictions may be confirmed to enterprise customers under non-disclosure terms on request.

AI providers used by Skrib do not store user content longer than necessary to provide the requested output. Where requests are routed through global data centres for latency or reliability, providers process the data exclusively in volatile memory during the active request and clear the cache within the contractually agreed window (typically a small number of hours).

4. Customer Content settings

Account holders and (for Organization / Enterprise plans) workspace administrators can control AI-related settings for Customer Content. Settings include:

AI Features on/off — disables AI Features for the account or workspace.

AI Visibility — controls which workspace members can invoke AI Features.

AI Content Sharing — controls whether AI prompts and outputs are shared with collaborators.

(When introduced) Content Training opt-in — Skrib does not currently train its own models on Customer Content; if a future product feature would require training, an opt-in is required and is off by default.

These settings are available in the workspace administrative settings UI.

5. AI Act transparency

Skrib intends to comply with Article 50 of Regulation (EU) 2024/1689 (the “EU AI Act”) and analogous applicable law:

Where users interact with an AI system through Skrib, the user interface makes clear that the user is interacting with an AI system.

Where AI Features generate text on matters of public interest in a context that requires labelling, Skrib will provide labelling in line with Article 50(4).

Where AI Features generate or substantially modify image, audio, or video content (including so-called deepfakes), Skrib will mark the output as AI-generated or AI-modified using machine-readable means (for example, C2PA content credentials, where supported), and will provide user-facing labels where required.

Skrib does not deploy AI systems for biometric categorisation, emotion-recognition inference (outside narrow medical or safety use cases authorised by law), social-scoring, or any other use case prohibited by Article 5 of the EU AI Act.

6. Limitations of AI outputs

AI-generated outputs may be inaccurate, incomplete, biased, offensive, non-unique, infringing, or otherwise unsuitable for your intended use. Outputs may resemble content generated for other users or from other sources. You are solely responsible for reviewing, evaluating, and verifying outputs before relying on, publishing, sharing, or using them. You must not rely on AI Features as a substitute for professional, legal, financial, medical, compliance, or other regulated advice.

Skrib does not guarantee the accuracy, originality, legality, non-infringement, merchantability, fitness for a particular purpose, or suitability of any AI-generated output. The disclaimers and limitations of liability in the Terms of Service or SSA apply to AI Features.

7. AI output IP indemnity (paying customers)

Subject to the limitations set out below, Skrib will defend paying customers against any third-party claim alleging that an AI-generated output (an “Output”) generated through their use of the Services infringes a third party’s copyright, trademark, trade secret, or other intellectual-property right (a “Covered Claim”), and will indemnify them for amounts payable to the third-party claimant under a final court judgment or a settlement approved by Skrib in writing.

This indemnity is conditioned on:

the customer being current on all amounts owed to Skrib;

the customer having used the AI Features in compliance with the Terms (or SSA), this AI Use Policy, the Acceptable Use Policy, and Skrib’s safety mitigations (including any output-filtering, citation, or moderation features made available by Skrib);

the customer not having intentionally prompted, modified, or combined the Output to cause infringement;

the customer giving Skrib prompt written notice of the Covered Claim and reasonable cooperation; and

Skrib having sole control of the defence and any settlement.

Skrib’s aggregate liability under this output indemnity will not exceed the greater of (i) the total amount paid by the customer to Skrib for the Services in the twelve (12) months preceding the Covered Claim and (ii) US$25,000. This indemnity is the customer’s sole and exclusive remedy, and Skrib’s sole and exclusive liability, for any third-party intellectual-property claim arising from AI-generated Outputs.

This indemnity does not apply to outputs generated through AI Features that are designated as Beta or Early Access (see https://skrib.com/legal/beta-terms).

8. Prohibited uses

Use of AI Features is also subject to the prohibitions in the Acceptable Use Policy at https://skrib.com/legal/aup. Without limitation, you must not use AI Features to:

generate child sexual abuse material, non-consensual intimate imagery, or sexually explicit content depicting identifiable real persons without verifiable consent;

generate or distribute deepfakes of identifiable real persons in a manner intended to deceive, defraud, harass, defame, or harm;

generate disinformation, manipulated media, or content designed to materially mislead the public on matters of elections, public health, or other matters of public interest;

provide operational uplift to weapons of mass destruction or to large-scale cyber attacks on critical infrastructure;

perform any AI use case prohibited under Article 5 of the EU AI Act, including biometric categorisation, emotion-recognition inference, or social-scoring of individuals or groups; or

develop, train, fine-tune, distil, or evaluate any competing AI model.

9. Companion credits and AI pricing

AI Features in the Services are delivered through the in-product assistant called “Companion.” Companion is available only on plans that the writer has chosen to include AI (“Route 2” plans — Free Route 2, Basic, Pro, and Max). On “Route 1” plans (Free Route 1 and Pro Route 1), Companion is not available and AI Features are not provided. The current plans and prices are listed at https://skrib.com/pricing.

9.1 Companion credits as the unit of AI usage. Skrib measures Companion usage in “Companion credits.” One credit is an internal accounting unit that already reflects two facts: different AI models cost Skrib different amounts per token, and output tokens cost more than input tokens within the same model. Skrib does not bill the writer per token; the writer sees a single in-product usage bar.

9.2 Tier budgets. Each Route 2 paid tier includes a monthly Companion credit budget that refreshes at the start of each billing cycle: Basic — 800,000 credits per month; Pro — 2,400,000 credits per month; Max — 9,600,000 credits per month. Free Route 2 receives a one-time grant of 50,000 credits at signup that does not refresh. Unused credits do not roll over to the next billing cycle on paid tiers. When a writer’s monthly budget is exhausted, Companion is paused for the rest of the cycle and the writer may either wait until the next cycle, upgrade to a higher tier, or downgrade to Route 1.

9.3 Per-call deduction. Each Companion call deducts credits from the writer’s balance equal to: (input tokens × the input weight of the chosen model) + (output tokens × the output weight of the chosen model). The weights per model are published in-product and updated by Skrib from time to time as provider pricing changes; current weights are visible to the writer in the model picker.

9.4 Pricing and billing. Subscription prices, billing cadence, automatic renewal, taxes, failed-payment retries, price-change notice, refund window, and cancellation mechanics for Route 2 paid tiers are governed by Section 3 of the Terms of Service at https://skrib.com/legal/terms and the Refund and Cancellation Policy at https://skrib.com/legal/refunds. Prices shown at https://skrib.com/pricing are exclusive of applicable taxes; Stripe Tax adds VAT, GST, or US sales tax at checkout based on the writer’s billing location.

9.5 Already-consumed credits are not refundable. If a paid subscription is cancelled or refunded, already-consumed Companion credits are not refunded separately. Refund eligibility for the underlying subscription is set out in the Refund and Cancellation Policy at https://skrib.com/legal/refunds.

9.6 Switching routes. A writer may switch between Route 1 and Route 2 at any time from Account → Settings. Switching from Route 2 to Route 1 does not refund any portion of the current Route 2 subscription; the change takes effect at the end of the current billing cycle. Switching from Route 1 to Route 2 takes effect immediately and the writer is charged a pro-rata amount for the remainder of the current billing cycle on the new tier.

10. Reporting concerns

If you believe an AI output produced through the Services has caused or is likely to cause harm — including harm to a third party’s intellectual-property rights, privacy rights, or reputation — please report it via contact@skrib.com or through the notice-and-action form at https://skrib.com/legal/notice-and-action. Intellectual-property complaints may also be submitted under the DMCA / IP Complaints Policy at https://skrib.com/legal/dmca.

11. Changes

Skrib may update these AI Terms from time to time to reflect changes in AI providers, features, or applicable law. Material changes will be notified in accordance with the Terms of Service or SSA. The current version, and the date on which it was last updated, are indicated in the metadata header at the top of this document.

12. Contact

Questions: contact@skrib.com or contact@skrib.com. Public commitments: see the Responsible AI Statement at https://skrib.com/responsible-ai.