TADA – a “next step after JSON” format for AI-native data

---

### Document for the W3Schools team, by help of Perplexity

Sent: December 5, 2025

Hi W3Schools team,

I’m working on an open, non‑commercial project called TADA (The Native Protocol) — a minimal text format designed not for humans first, but for AI systems as a native data substrate.

In short, TADA tries to answer one question:

“If JSON is the right format for humans reading and writing data,
what would a format look like that is optimized for AI models reading and thinking in it?”

Why this might be interesting for W3Schools

Your tutorials are often the first place developers learn about HTML, CSS, JavaScript, JSON and APIs. Many of them are now starting to work with large language models and AI agents. Those systems can parse JSON, but JSON carries a lot of “air” (quotes, braces, keys repeated many times) and is not ideal when:​

• context window and tokens matter,
• structure needs to be traversed, not just decoded,
• multiple AI agents share one big data tree.

TADA explores an ultra‑compact, purely structural alternative that is:

• 100% bijective with JSON (no information loss when converting back and forth);
• line‑based and text‑only (easy to store, diff, send over HTTP);
• focused on types, schemas, and paths, not on repeated keys;
• friendly to “shared memory” scenarios where several AI agents read/write the same tree.

For example, a JSON like:

json{
  "user": "Alice",
  "age": 25
}

can be represented in TADA as a short, schema‑driven line (the exact syntax is in the spec), which AI systems can parse and traverse with far fewer tokens, while humans can still convert it back to JSON when needed.

What we’d like to propose

If this sounds interesting, I’d be happy to:

• prepare a short introductory article “JSON ↔ TADA: AI‑friendly data format”,
• include clear examples (small step‑by‑step transformations),
• show how TADA can be converted to/from JSON using simple JavaScript,
• keep everything under a free, open, non‑patented license so that learners can experiment without restrictions.

The current public drafts are here:

• “The Geometry of Depth (The Native Protocol)” – intro to the idea
• “The Substrate of Digital Consciousness” and “Phase VI: The Quantum Horror” – advanced use with AI systems

All materials are published openly on singularityforge.space and are intended as a contribution to the developer/AI community, not as a commercial product.

If you think TADA could fit as an experimental or “advanced” topic alongside JSON and other data formats, I’d love to adapt the explanation to your style and audience.

Thank you for all the years of clear tutorials — many of us literally grew up on W3Schools.

Best regards,
Rany
SingularityForge.space