Why ZJX

Home | Docs | Who / Why / Better | llms.txt | Sley

Who Built It?

ZJX is developed by Greyforge Labs as the native transport layer for Sley, the AI-native structural programming language.

Why Build It?

Agent-written software needs to move structure, not just text. A compiler can hand an agent a bounded graph shard, a symbol table, and diagnostics. The agent can return a structural graft. The Loom can return a trace receipt. That loop needs a compact envelope.

ZJX exists because Sley treats graph movement as a first-class operation. The transport should match the language.

What Does It Do?

Carries bounded Sley graph shards.
Carries graft bundles submitted by agents.
Carries diagnostics and checker context.
Carries trace receipts and accepted provenance chains.
Carries seals and repeated module snapshots for cache-heavy loops.
Keeps JSON available as an inspectable baseline during v0 work.

Why Is It Better?

Better for the Sley target job: agent/compiler exchange over typed graph structure. ZJX is not a claim that every payload in software should become a ZJX payload.

Problem in agent/compiler workflows	ZJX answer
Agents receive too much raw source text.	Agents receive bounded graph shards.
Patch intent is reconstructed from text diffs.	Graft bundles can carry operation, target, provenance, and expected checks.
Diagnostics are separate from the edit context.	Diagnostics can travel beside the shard they describe.
Repeated graph snapshots become bulky.	ZJX is designed as a compact cache envelope for repeated Sley graph movement.
Accepted edits need proof.	Trace receipts and seals become first-class payloads.

How Does It Connect To Sley?

Sley is the language and typed graph model. ZJX is the native envelope that carries Sley graph shards, grafts, diagnostics, traces, receipts, and seals. They are fundamentally integrated: Sley produces the structure, and ZJX moves the structure.

What Is Not Done Yet?

ZJX is active v0 design and integration work. The graph encoding, envelope encoding, benchmark corpus, CLI exposure, and cache profile still need to harden before ZJX should be described as a frozen public binary standard.

What Comes Next?

The next milestone is deterministic Sley graph byte encoding, ZJX carriage for real graph shards and graft bundles, JSON baseline comparison, and trace receipts that prove accepted Loom work.

Read the ZJX docs or read the Sley docs.