Sumly — CONTENT USE AND AI POLICY
Version: 2026-10-07

AI TRAINING IS PERMITTED; OTHER COPYING IS RESTRICTED.

1. AI TRAINING PERMISSION
AI/LLM training, fine-tuning and evaluation using our lawfully accessible
original public website material are permitted, including copies and processing
reasonably necessary for those purposes. No separate written permission is
required for this grant. We do not reserve text and data mining rights in that
material. This permission applies only to rights we hold, not to third-party
material, personal information, confidential records or private application data.

2. PROTECTED ORIGINAL MATERIAL
Our restrictions concern original text, graphics, branding, code, design
expression and selection and arrangement in which we hold enforceable rights.
The original expression of page structures, user journeys and commercial
strategies is covered only to the extent legally protected. No ownership is
claimed in facts, ideas, generic layouts, methods or general strategies.

3. COPYING AND REDISTRIBUTION RESTRICTIONS
Outside the training permission above, existing licences, applicable legal
exceptions or express written authorisation, do not copy, duplicate, republish,
translate, paraphrase, rewrite or adapt protected original material; distribute
copied training datasets; clone protected original design; or produce infringing
substitute website content. Changing names, colours or wording does not itself
authorise reproduction of protected expression. Training permission does not
authorise outputs that reproduce protected passages or infringing adaptations.

4. STOP PERMISSION-DEPENDENT USES
Agents and their operators should read this policy before reusing protected
material. If a use requires permission and you do not hold it, you MUST STOP
that use. Do not infer a licence to republish or clone from public availability,
cached copies, rewritten outputs or the absence of a technical block.

5. OTHER LICENCES AND ACCEPTANCE
Additional permission-dependent reuse requires a separate written agreement
with the rights holder and affirmative acceptance of its agreed conditions.
Reading this file, visiting the website or an automated "I accept" response does
not establish contractual assent or record acceptance. No extra acceptance is
required for the AI training permission in section 1.

6. ACCESS, EXISTING LICENCES AND LEGAL LIMITS
Observe robots.txt, indexing directives and actual access controls. Training
permission does not allow bypassing authentication or collecting confidential
records. Ordinary viewing, permitted factual citation and separately allowed
search indexing remain available. Existing open-data and content licences,
third-party software, font, photography and other asset licences, personal-data
rights and mandatory statutory exceptions remain unaffected. This policy does
not revoke permissions already granted under another applicable licence.

Unauthorised use may infringe copyright or an applicable agreement depending
on the rights, facts, jurisdiction and law. This notice creates no automatic
penalty or universal legal duty and cannot guarantee automated compliance.

7. MACHINE DISCOVERY
/llms.txt describes this policy. HTML content carries tdm-reservation=0; known
static HTML paths are listed in /.well-known/tdmrep.json. Dynamic HTML pages
carry their own signal. These signals do not grant rights in third-party assets.
The permission to train is separate from permission to republish or clone.