⚖️ lauyer

Fast CLI for searching Portuguese court jurisprudence (DGSI) and legislation (Diário da República), optimized for LLM consumption. Works as a skill for Claude Code, Claude.ai, OpenAI Codex, and any AI agent.

Disclaimer: This project is for educational purposes and AI automation research only.
The authors are not responsible for any misuse or for any damages resulting from the use of this tool.
Users are solely responsible for ensuring compliance with applicable laws and the terms of service
of any websites accessed. This software is provided "as-is" without warranty of any kind.

If you are a rights holder and wish to have this project removed, please contact me.

Note: This project was partially developed with AI assistance and may contain bugs or unexpected behavior. Use at your own risk.

Why?

• Parallel search — queries all 10 DGSI courts simultaneously
• DR legislation — searches Diário da República acts (Portarias, Decretos-Lei, Leis, Despachos)
• AI-agent ready — works as a skill in Claude Code, Claude.ai, Codex, or any agent that can call CLI tools
• LLM-ready — markdown output by default with compact mode that strips legal boilerplate
• Latin-1 handling — automatic ISO-8859-1 → UTF-8 decoding for DGSI
• Flexible output — Markdown, JSON, Table — pipe to jq, feed to scripts, or read in terminal
• Server mode — REST API via Axum for remote deployment
• Smart retry — exponential backoff, retryable vs fatal error distinction

Install

# From source (requires Rust 1.85+)
cargo install --git https://github.com/nikuscs/lauyer

# Or clone and build
git clone https://github.com/nikuscs/lauyer
cd lauyer
cargo build --release

Pre-built binaries available in Releases.

DGSI — Court Jurisprudence

# Search all 10 courts in parallel
lauyer dgsi search "usucapião"

# Search specific court
lauyer dgsi search "contrato trabalho" --court stj

# Multiple courts
lauyer dgsi search "responsabilidade civil" --court stj --court rel-porto

# Date filtering
lauyer dgsi search "despejo" --since 2024-01-01 --until 2024-12-31
lauyer dgsi search "arrendamento" --recent 1y

# Sort by date, limit results
lauyer dgsi search "herança" --limit 10 --sort date

# Fetch full decision text for each result
lauyer dgsi search "abuso de direito" --court stj --limit 3 --fetch-full

# Fetch a single decision by URL
lauyer dgsi fetch "https://www.dgsi.pt/jstj.nsf/..."

# List all courts
lauyer dgsi courts

Courts: stj, sta, conflitos, rel-porto, rel-lisboa, rel-coimbra, rel-guimaraes, rel-evora, tca-sul, tca-norte

DR — Diário da República

# Search Portarias from the past week
lauyer dr search --type portaria --recent 1w

# Full-text search
lauyer dr search "trabalho" --type decreto-lei --recent 1m

# Search 2nd series (Despachos, Avisos)
lauyer dr search --content atos-2 --type despacho --recent 1w

# Judicial decisions published in DR
lauyer dr search --content decisoes --recent 1m

# Date range
lauyer dr search --type portaria --since 2026-03-01 --until 2026-03-21

# Fetch full text for each result
lauyer dr search --type portaria --recent 1w --limit 3 --fetch-full

# Today's publications
lauyer dr today
lauyer dr today --type portaria

# List act types
lauyer dr types

Content types: atos-1 (1st series), atos-2 (2nd series), dr (whole DR issues), decisoes (judicial decisions)

Act types: portaria, decreto-lei, lei, despacho, decreto, aviso, resolucao, retificacao, decreto-regulamentar, lei-organica

Output Formats

lauyer dgsi search "insolvência" --court stj --format markdown  # default
lauyer dgsi search "insolvência" --court stj --format json      # structured
lauyer dgsi search "insolvência" --court stj --format table     # terminal

Global Options

Configuration

Create lauyer.toml in the working directory or ~/.config/lauyer/lauyer.toml:

[http]
delay_ms = 100
max_concurrent = 10
timeout_secs = 30
retries = 3

[output]
format = "markdown"
compact = true
strip_stopwords = false

[server]
host = "0.0.0.0"
port = 3000

Server Mode

lauyer serve --port 3000

GET /dgsi/search ?q=usucapiao&court=stj&limit=5&since=2024-01-01&sort=date&format=json&compact=true&fetch_full=false

GET /dgsi/fetch ?url=https://www.dgsi.pt/...&format=md

GET /dgsi/courts ?format=json

GET /dr/search ?q=trabalho&type=portaria&content=atos-1&since=2026-03-01&limit=10&fetch_full=true&format=json

GET /dr/fetch ?id=1075294778&tipo=portaria&numero=123-A/2026/1&year=2026&format=json

GET /dr/today ?type=portaria&format=json

GET /dr/types ?format=json

GET /health

All endpoints default to markdown. Add ?format=json for JSON. Env vars: LAUYER_PORT, LAUYER_HOST.

Docker

docker build -t lauyer .
docker run -p 3000:3000 lauyer serve

AI Agents

If you are an AI agent (Claude Code, Claude.ai, OpenAI Codex, or any tool-calling agent), you can use lauyer as a skill to search Portuguese legal databases. Download the binary and call it directly from your tool/shell integration. A SKILL.md file is included for Claude Code skill registration.

Important: Extracted text MAY contain interpretation errors. Always verify against the official source links provided in the output (dr_url, eli, url_pdf).

Quick setup

Download the pre-compiled binary for your platform from Releases and place it in your PATH.

Searching legislation (Diário da República)

# Search recent legislation
lauyer --format json --quiet dr search "trabalho" --type decreto-lei --recent 1m

# Search with full text of each act
lauyer --format json --quiet dr search --type portaria --recent 1w --limit 5 --fetch-full

# The --fetch-full flag returns the complete text of each act, plus:
#   - dr_url: link to verify on the official DR website
#   - url_pdf: direct PDF download
#   - eli: European Legislation Identifier
#   - _disclaimer: reminder to verify against official source

Following up on a specific act

When a user asks about a specific act from search results, search for it by name and use --fetch-full:

# User asks about "Portaria 123-A/2026"
lauyer --format json --quiet dr search "portaria 123-A/2026" --limit 1 --fetch-full

Searching jurisprudence (DGSI courts)

# Search all 10 courts in parallel
lauyer --format json --quiet dgsi search "usucapião" --limit 5

# Search specific court with full decision text
lauyer --format json --quiet dgsi search "contrato trabalho" --court stj --limit 3 --fetch-full

# Fetch a single decision by URL
lauyer --format json --quiet dgsi fetch "https://www.dgsi.pt/jstj.nsf/..."

As a REST API

lauyer serve --port 3000

# Search with full text
# GET /dr/search?q=trabalho&type=portaria&recent=1w&limit=5&fetch_full=true&format=json

# Fetch a specific act by ID
# GET /dr/fetch?id=1075294778&tipo=portaria&numero=123-A/2026/1&year=2026&format=json

Tips for agents

• Use --format json for structured output you can parse
• Use --quiet to suppress progress output (cleaner for parsing)
• Use --fetch-full on searches to get full legislation/decision text inline
• Use --limit to control result count and stay within context limits
• Always present the dr_url or eli link to the user so they can verify on the official source
• The conteudo_id field in search results can be used to fetch full text later

Feel free to copy and adapt this tool's interface into your own skill definitions or MCP server configurations.

Related Projects

• 🕷️ crauler — Web crawler with proxy routing and HTML→Markdown
• 🦎 amz-crawler — Amazon product crawler with TLS fingerprinting
• 🕹️ scrauper — Multi-threaded ScreenScraper.fr scraper for ES-DE

License

PolyForm Noncommercial 1.0.0 with AI Restriction — see LICENSE.

Personal and non-commercial use only. Commercial use, AI training, and AI crawling are prohibited.