Fox Valley AI Foundation

§ CTX

Context Transfer Format — the content layer between the web and AI
87–90% token reduction on real-world web pages

GitHub · Spec v1.0 · foxfoundation.ai · @NicW_AI

→ Input

Web page URL

Extraction Tier

fast = DOM rules only (<500ms) · smart = DOM + regex NER (<1s)

fast smart

→ What is CTX?

§doc.ctx_v1.0 url=example.com/article title="Article" †type=article †lang=en §nav [skip] Main menu... §content.article §1 Introduction §p Content with citations [ref1] preserved... §2 Subsection §p Hierarchy intact, zero wasted tokens. §footer [skip] Copyright... §ref id=ref1 url=example.com title="Source" †rel=related

Fetch → Extract → Classify → Annotate → Normalize → Emit → § CTX
Full spec-compliant pipeline · readability + BeautifulSoup DOM · inline citations · skip annotations · section nesting

→ Results

§ CTX Output

        Waiting for input...
    

→ Try these

Web page URL	Raw HTML	Extraction Tier