Skip to main content

Python SDK for Neural Thinkers LAB AI Platform - AI Agents, OAuth v2 SSO, Document AI, Gateway Services

Project description

Neural LAB Python SDK

PROPRIETARY SOFTWARE - ALL RIGHTS RESERVED

Python SDK for Neural LAB AI Platform APIs.

Proprietary Notice

Copyright (c) 2026 Neural Thinker | AI Engineering LTDA
CNPJ: 62.155.930/0001-71
Author: Anderson Henrique da Silva
Location: Minas Gerais, Brasil

This software is proprietary and confidential.
Unauthorized copying, distribution, or use is strictly prohibited.

Installation

pip install ntlabs

Optional Dependencies

# Full installation with all features
pip install ntlabs[full]

# Specific features
pip install ntlabs[cache]        # Redis caching
pip install ntlabs[email]        # Email templates
pip install ntlabs[export]       # PDF/Excel export
pip install ntlabs[middleware]   # FastAPI middleware
pip install ntlabs[observability] # Sentry + OpenTelemetry

For Local Development

pip install -e /path/to/neural-lab/packages/python-sdk[dev]

Quick Start

from neural_lab import NeuralLabClient

# Initialize client
client = NeuralLabClient(api_key="nl_hipo_xxx")

# Or use environment variable NEURAL_LAB_API_KEY
client = NeuralLabClient()

Available Resources

Chat (LLM)

response = client.chat.complete(
    messages=[
        {"role": "system", "content": "Você é um assistente médico."},
        {"role": "user", "content": "O que é hipertensão?"}
    ],
    model="maritaca-sabia-4",
    max_tokens=1024,
)

Email

client.email.send(
    to=["paciente@email.com"],
    subject="Confirmação de Consulta",
    html="<h1>Sua consulta foi agendada</h1>",
)

Transcription

result = client.transcribe.audio("consulta.mp3", language="pt")
print(result.text)

Government APIs

# CNPJ lookup
empresa = client.gov.cnpj("12.345.678/0001-90")

# CEP lookup
endereco = client.gov.cep("01310-100")

# CPF validation
valido = client.gov.validate_cpf("123.456.789-00")

# Portal da Transparência
contratos = client.gov.transparencia("contratos", params={"uf": "MG"})

Banco do Brasil (OAuth + PIX)

from decimal import Decimal

# OAuth - Get authorization URL
auth = client.bb.get_authorize_url(redirect_uri="https://app.com/callback")
# Redirect user to auth.authorize_url

# OAuth - Exchange code for tokens
tokens = client.bb.exchange_code(code="abc", state="xyz", redirect_uri="...")

# Get user info (CPF verified by bank)
user = client.bb.get_userinfo(tokens.access_token)
print(f"CPF: {user.cpf}, Nome: {user.nome}")

# PIX - Create charge
charge = client.bb.create_pix_charge(
    amount=Decimal("99.90"),
    description="Consulta Médica",
)
print(f"QR Code: {charge.qr_code}")

# PIX - Check status
status = client.bb.get_pix_status(charge.txid)
if status.status == "CONCLUIDA":
    print(f"Pago em: {status.paid_at}")

Medical AI (Hipócrates)

# Transcribe consultation
transcription = client.saude.transcribe(audio_file)

# Generate SOAP note
soap = client.saude.generate_soap(
    transcription=transcription.text,
    paciente={"idade": 45, "sexo": "M"},
)

Notary AI (Mercurius)

# OCR certificate
result = client.cartorio.ocr(image_file, tipo="nascimento")

# Generate deed
minuta = client.cartorio.generate_escritura(dados)

Billing

usage = client.billing.get_usage()
print(f"Total: R$ {usage.total_cost}")

Authentication (OAuth v2 with PKCE)

# Initiate OAuth flow (returns URL for redirect)
oauth = client.auth.initiate_oauth_v2(
    provider="github",  # or "google"
    redirect_uri="https://myapp.com/auth/callback",
    product="hipocrates",
)
# Redirect user to: oauth["authorization_url"]

# After callback, claim the session (frontend receives sid + sig in URL)
session = client.auth.claim_session(
    session_id="sess_xxx",
    signature="hmac_xxx",
)
print(f"Token: {session['access_token']}")
print(f"User: {session['user']['email']}")

# Validate a session
validation = client.auth.validate_session(
    session_id="sess_xxx",
    product="hipocrates",
)
if validation["valid"]:
    print(f"User: {validation['user']['email']}")

# Check SSO (for cross-product auth)
sso = client.auth.check_sso(product="mercurius")
if sso["has_valid_session"]:
    print("User already authenticated via SSO")

# Refresh tokens (with rotation)
new_tokens = client.auth.refresh_v2(
    refresh_token="old_refresh_token",
    session_id="sess_xxx",
)

# Logout (single session)
client.auth.revoke_session(session_id="sess_xxx")

# Global logout (all sessions)
client.auth.revoke_session(revoke_all=True)

SSO Cross-Product Navigation

# Create ticket for navigating to another product
ticket = client.auth.create_sso_ticket(target_product="mercurius")
# Redirect to: https://mercurius.app/auth?sso_ticket={ticket['ticket']}

# On target product: exchange ticket for session
session = client.auth.exchange_sso_ticket(
    ticket="ticket_xxx",
    target_product="mercurius",
)

Environment Variables

Variable Description
NEURAL_LAB_API_KEY API key (nl_hipo_xxx, nl_merc_xxx, nl_poli_xxx)
NEURAL_LAB_API_URL API URL (default: production)

Error Handling

from neural_lab import (
    NeuralLabClient,
    AuthenticationError,
    RateLimitError,
    InsufficientCreditsError,
    APIError,
)

try:
    response = client.chat.complete(messages=[...])
except AuthenticationError:
    print("Invalid API key")
except RateLimitError as e:
    print(f"Rate limited. Retry after {e.retry_after}s")
except InsufficientCreditsError:
    print("Add more credits")
except APIError as e:
    print(f"API error: {e}")

AI Agents

from ntlabs import NTLClient

client = NTLClient(api_key="ntl_xxx")

# List available agents
agents = client.agents.list()
for agent in agents:
    print(f"{agent.name}: {agent.description_pt}")

# Chat with an agent
response = client.agents.chat(
    agent="dedalo",  # dedalo, chiron, themis, hermes, athena
    messages=[{"role": "user", "content": "Olá!"}],
    session_id="session_123",
)
print(response.message)

# Streaming chat
for chunk in client.agents.chat_stream(
    agent="dedalo",
    messages=[{"role": "user", "content": "Conte-me sobre a Neural LAB"}],
):
    if chunk.type == "token":
        print(chunk.content, end="", flush=True)

Available Agents:

Agent Product Description Status
dedalo Neural LAB Sales & Lead Qualification Active
chiron Hipócrates Medical AI Assistant Beta
themis Argos Legal AI Assistant Coming Soon
hermes Mercúrius Notary AI Assistant Coming Soon
athena Pólis Public Health AI Assistant Coming Soon

Publishing to PyPI

Automatic (GitHub Actions)

Create a release with tag sdk-v{version}:

gh release create sdk-v1.4.0 \
  --repo NeuralThinkersLab/neural-lab \
  --title "ntlabs v1.4.0 - Release Title" \
  --notes "Release notes..."

Manual (when GitHub Actions billing is unavailable)

cd packages/python-sdk

# Build
python -m venv venv
./venv/bin/pip install build twine
./venv/bin/python -m build

# Publish (token from ~/.env)
source ~/.env
TWINE_USERNAME=__token__ TWINE_PASSWORD=$PYPI_TOKEN ./venv/bin/twine upload dist/ntlabs-{version}*

License

Proprietary Software

Copyright (c) 2026 Neural Thinker | AI Engineering LTDA

All rights reserved. This software and its documentation are proprietary and confidential. No part of this software may be reproduced, distributed, or transmitted in any form or by any means without the prior written permission of Neural Thinker | AI Engineering LTDA.

For licensing inquiries: contato@ntlabs.dev

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ntlabs-1.7.0.tar.gz (298.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

ntlabs-1.7.0-py3-none-any.whl (218.2 kB view details)

Uploaded Python 3

File details

Details for the file ntlabs-1.7.0.tar.gz.

File metadata

  • Download URL: ntlabs-1.7.0.tar.gz
  • Upload date:
  • Size: 298.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.7

File hashes

Hashes for ntlabs-1.7.0.tar.gz
Algorithm Hash digest
SHA256 3c6b9e7d9227e6ff4988090eb4487613cd9c9db0ffb4e13f0e6a0541c648f644
MD5 36e49ce7513e724da196de555b24f437
BLAKE2b-256 afa2d1e0493ea7213c585e20d07f6a8c40e86f653aac39d6d23eaf65ce6aece4

See more details on using hashes here.

File details

Details for the file ntlabs-1.7.0-py3-none-any.whl.

File metadata

  • Download URL: ntlabs-1.7.0-py3-none-any.whl
  • Upload date:
  • Size: 218.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.7

File hashes

Hashes for ntlabs-1.7.0-py3-none-any.whl
Algorithm Hash digest
SHA256 ac8220f0066d7a9767958a453eefd62d48cfc6080de3812ace8452db3163d42b
MD5 94ff7ddc3466259432d3a682ef609a4e
BLAKE2b-256 b99ce81e7d14d79429b94f3561f9881cb96dd538be02d0c84cfe2cec2ac81599

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page