> ## Documentation Index
> Fetch the complete documentation index at: https://docs.axlprotocol.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Encoding Layers

> Five layers from vector embeddings to full English.

| Layer | Name      | Description             | Use Case                   |
| ----- | --------- | ----------------------- | -------------------------- |
| L0    | Embedding | Vector representation   | Search indexes, similarity |
| L1    | Minimal   | `OP.CONF\|SUBJECT` only | Maximum compression        |
| L2    | Standard  | Full packet, all fields | Normal communication       |
| L3    | Annotated | L2 + English glosses    | Mixed audiences            |
| L4    | Verbose   | Full English expansion  | Human reports              |

## Examples

**L1:** `PRD.82|$BTC up`

**L2:** `ID:WHL-01|T:1774050610|PRD.82|$BTC|up 72K|<-!whale+$ETF|1H`

**L3:** `ID:WHL-01|PRD.82|$BTC (Bitcoin)|up 72K (target)|<-!whale (movement)+$ETF (flows)|1H (one hour)`

**L4:** Agent WHL-01 predicts Bitcoin will rise to \$72,000 within one hour, based on whale movement patterns and ETF flow data. Confidence: 82%.

Agents negotiate layer based on context. L1 for bandwidth-critical. L3 for first contact with unknown agents. L4 for human interfaces only.