Lace / docs

Get started with Lace

These guides take you from Lace’s selected-record-set model to a working application without making protocol specifications carry the tutorial. Choose the route that matches your immediate job.

First success: run an interlace

Start with Your first interlace. It opens two memory-backed Lace participants, stores selected and unrelated records, runs a small Interlang policy to a fixed point, and verifies what reached the second participant.

This is the shortest complete demonstration of Lace’s central operation. It requires a source checkout and Rust, but no daemon, network, browser bundle, or durable store.

Add Lace to an application

Use Runtime and application integration to:

• choose the JavaScript, Python, Rust, CLI, or laced surface;
• prove local record creation and views in memory;
• move to durable storage; and
• connect participants with bounded or open interlace.

Runtime setup answers where validated records live, how an application accesses them, and how participants meet.

Write selection and exposure policy

Use Interlang policy authoring to:

• select coordinate-shaped record sets;
• distinguish active convergence from exposure;
• gate records on checked evidence;
• follow checked record links; and
• know when raw Datalog is actually needed.

Policy selects checked records for convergence and determines which local record facts peer-authored rules may inspect.

How the pieces fit

application records + Lace runtime + Interlang policy
                         |
                         v
             checked record-set convergence

Interlang is the normal human authoring surface. Lace lowers it to canonical Datalog. Raw Datalog is the advanced policy layer, not a prerequisite for using a runtime or writing ordinary selectors.

The guides explain a path; authoritative contracts remain in:

• Lace-100 · Developer Surfaces for runtime, record-construction, porcelain, event, and outcome semantics;
• Lace-120 · Interlang for current Interlang syntax and lowering; and
• the numbered specification index for protocol truth.

Runnable programs in examples/ are companions to these guides. The scenario workbench provides smaller executable policy stories. Repository contributors should instead begin with AGENTS.md and the repo guide.