Treeia-Token is an AI-native symbolic format, automatically derived from the Treeia-STS binary. It encodes structures, parameters, and primitive types as PUA tokens, with no human text and no punctuation. Its grammar is closed and schema-driven, enabling linear-time validation.
With textual formats (JSON, XML, free prompts), an AI must:
- infer structure
- interpret roles
- manage syntax
- keep large context in memory
This introduces ambiguity, structural errors, and structual hallucinations.
Treeia-Token provides a deterministic token stream where all structure is dictated by STS schemas, not inferred from syntax:
- closed grammar
- schema-driven typing
- positional decoding
- bijection with STS
- STS-derived tokens — every struct, parameter, and primitive type maps to a unique symbol.
- Linear symbolic flow — no words, no punctuation, no syntactic choices.
- Self-delimited blocks — clear structure, trivial parsing.
The AI no longer guesses. It reads, generates, and validates a constrained token stream.
Treeia separates container from geometry.
tispi is the internal geometry — the Time/Space structure itself (Piece / Face, Timeline, Tracks, Keys).
In short:
- treeiaj = transport layer
- tispi = structural engine
The container is generic. The geometry defines meaning.
- Total determinism
- Zero text, zero ambiguity
- AI-native (short tokens, closed grammar)
- Extensible by construction (new structs → new tokens automatically)
Types:
{
"Coord": { "x": "number", "y": "number" }
}Script:
{ "Coord": { "x": 100, "y": 200 } }In Treeia-Token:
→ a pure PUA token stream derived from the Treeia binary —
| Token | Signification (d’après la table STS) |
|---|---|
| ⟦S1⟧ | struct Coord (ID 1) |
| ⟦P0⟧ | premier paramètre de Coord (x) |
| ⟦Tfloat⟧ | le type de la valeur est float |
| 100 | valeur brute du float (little‑endian) |
| ⟦P1⟧ | second paramètre de Coord (y) |
| ⟦Tfloat⟧ | type float |
| 200 | valeur |
no words, no punctuation, only symbols and raw values.
Treeia‑Token represents an exploration of how to eliminate ambiguity in AI data exchange. It is not yet a product, but a framework for discussion.