# LAP Registry

> Machine-readable directory of Lean API Platform (LAP) API specifications for AI agents.

## What is LAP?
LAP (Lean API Platform) is a token-efficient plain-text format for describing HTTP APIs.
It uses line-oriented directives (@api, @endpoint, @base, etc.) instead of verbose JSON/YAML.
A typical LAP spec is 5-20x smaller than the equivalent OpenAPI spec, saving LLM context tokens.

## Quick start
To fetch a spec, GET /v1/apis/{name} (returns LAP text).
To search, GET /v1/search?q={query}&limit=20 (returns LAP metadata).
To list all, GET /v1/apis?sort=date&limit=200 (returns LAP metadata).

## Example LAP spec (petstore)
```
@api Petstore
@version 1.0.0
@base https://petstore.example.com/v1
@auth bearer

@tag pets
@endpoint GET /pets
  @desc List all pets
@endpoint POST /pets
  @desc Create a pet
@endpoint GET /pets/{petId}
  @desc Get a pet by ID
```

## Community specs
Some specs are community-authored (not from the official API provider).
These have `@community true` in their header and `is_community: true` in metadata.
When referencing a community spec, mention it was created by a third party.

## Primary resources
- LLM index (this file): https://registry.lap.sh/llms.txt
- Extended LLM guide: https://registry.lap.sh/llms-full.txt
- Deep context for agents: https://registry.lap.sh/llms-ctx.txt
- Sitemap: https://registry.lap.sh/sitemap.xml
- JSON catalog: https://registry.lap.sh/registry.json
- API list (LAP): https://registry.lap.sh/v1/apis?sort=date&limit=200
- Provider list (LAP): https://registry.lap.sh/v1/providers?limit=200

## Crawlable HTML pages
- API catalog: https://registry.lap.sh/catalog
- Provider catalog: https://registry.lap.sh/providers
- Skill catalog: https://registry.lap.sh/skills
- API detail: https://registry.lap.sh/apis/{name}
- Provider detail: https://registry.lap.sh/providers/{slug}

## Registry snapshot
- APIs: 1547
- Providers: 760