LAP CLI

Compile, search, and manage API specs for AI agents.

Command Resolution

Detect the available CLI binary once per session:

# Prefer global install, fall back to npx
if command -v lapsh &>/dev/null; then
  LAP_CMD="lapsh"
else
  LAP_CMD="npx @lap-platform/lapsh"
fi

Use $LAP_CMD (or the resolved form) for all commands below.

Agent Flow -- Consuming APIs

Use this flow when a user needs to find, download, or work with an API.

1. Discover

lapsh search <query> [--tag <tag>] [--sort relevance|popularity|date] [--limit <n>]

Search the LAP registry. Results show name, endpoint count, compression ratio, and a [skill] marker for installable skills.

# Find payment APIs sorted by popularity
lapsh search payment --sort popularity

# JSON output for scripting
lapsh search payment --json | jq '.results[] | select(.has_skill) | .name'

2. Acquire

Option A -- Install a skill (if [skill] marker present):

lapsh skill-install <name>
# Installs to ~/.claude/skills/<name>/

Option B -- Download the spec:

lapsh get <name> -o <name>.lap
lapsh get <name> --lean -o <name>.lean.lap

Option C -- Compile a local spec:

lapsh compile <spec-file> -o output.lap
lapsh compile <spec-file> -o output.lean.lap --lean

Supported formats: OpenAPI (YAML/JSON), GraphQL (SDL), AsyncAPI, Protobuf, Postman, Smithy. Format is auto-detected.

3. Use

Once you have a .lap file, read it directly -- LAP is designed for AI consumption. Key markers:

Marker	Meaning
`@api`	API name, version, base URL
`@endpoint`	HTTP method + path
`@param`	Parameter (query, path, header)
`@body`	Request body schema
`@response`	Response status + schema
`@required`	Required fields
`@error`	Error response

Publisher Flow -- Publishing APIs

Use this flow when a user wants to compile, package, and publish an API spec.

1. Authenticate

lapsh login
lapsh whoami        # verify

2. Compile

lapsh compile spec.yaml -o spec.lap
lapsh compile spec.yaml -o spec.lean.lap --lean

3. Generate Skill

# Basic skill (Layer 1 -- mechanical)
lapsh skill spec.yaml -o skills/ --no-ai

# AI-enhanced skill (Layer 2 -- requires claude CLI)
lapsh skill spec.yaml -o skills/ --ai

# Generate and install directly
lapsh skill spec.yaml --install

4. Publish

lapsh publish spec.yaml --provider stripe.com
lapsh publish spec.yaml --provider stripe.com --name charges --source-url https://...
lapsh publish spec.yaml --provider stripe.com --skill          # include skill
lapsh publish spec.yaml --provider stripe.com --skill --skill-ai  # with AI skill

5. Verify

lapsh search <name>   # confirm it appears in registry

6. Batch Operations

# Generate skills for all specs in a directory
lapsh skill-batch specs/ -o skills/

Quick Reference

Core Commands

Command	Description
`compile <spec>`	Compile API spec to LAP format

Registry Commands

Command	Description
`search <query>`	Search the LAP registry
`get <name>`	Download a LAP spec
`publish <spec>`	Compile and publish to registry
`login`	Authenticate via GitHub OAuth
`logout`	Revoke token
`whoami`	Show authenticated user

Skill Commands

Command	Description
`skill <spec>`	Generate Claude Code skill from spec
`skill-batch <dir>`	Batch generate skills
`skill-install <name>`	Install skill from registry

Error Recovery

Problem	Fix
`command not found: lapsh`	`npm install -g @lap-platform/lapsh` or use `npx @lap-platform/lapsh`
`Not authenticated`	Run `lapsh login` first
`Format detection failed`	Pass `-f openapi` (or graphql, asyncapi, protobuf, postman, smithy)
`403 Forbidden` on get	Spec may be private or registry may block without User-Agent -- update lapsh
`YAML parse error`	Check spec is valid YAML/JSON -- use a linter first
`Layer 2 requires claude`	Install Claude CLI or use `--no-ai` for Layer 1 skills
`Provider required`	`publish` needs `--provider <domain>` (e.g., `--provider stripe.com`)

Environment Variables

Variable	Purpose	Default
`LAP_REGISTRY`	Override registry URL	`https://registry.lap.sh`

References

{baseDir}/references/agent-flow.md -- Extended agent workflow with examples
{baseDir}/references/publisher-flow.md -- Extended publisher workflow with examples
{baseDir}/references/command-reference.md -- Complete command reference with all flags

lap

Safety Notice

Copy this and send it to your AI assistant to learn