Projects

Large systems, in production.

A sample of what the team has built and runs: platforms with hundreds of screens, thousands of tests, and AI where it makes a difference. The numbers come straight from the repositories.

7,000+
commits in 2026: 900k+
lines of code: 14k+
automated tests: 8
systems on this page

The portfolio

One foundation, eight systems.

Estação ClínicaHealthcare · Medical practice

AstroMatConRetail · Building materials

SentinelaOccupational health · Auditing

JundoSMBs · Marketing, sales and customer service

Quick foundation

Security, logging, agents and learning

Ideia EstruturadaRequirements gathering

AceleraMatConRetail network · Tax and BI

OrgSense AIHR and OHS · Psychosocial risks

Gestor360Management · Multiple businesses

In production or in use
Rolling out or ready for pilot

Featured case · Healthcare · Medical practice

Estação Clínica

See patients face to face. The note comes out ready for you to review.

The practice's platform: scheduling and patient intake over WhatsApp, lab result reading, a consultation scribe and a clinical copilot. Every sentence the AI writes points to where it came from, and nothing reaches the patient without the doctor's sign-off.

Visit the product website See the full case

Inside, in twelve screens. The real system, with a fictitious patient and lab results. A minute and a half, no sound.

What the doctor can count on

Names, tax IDs and phone numbers become codes before anything reaches the AI
Every suggestion shows its source
The consultation audio never leaves the server
Data stored in Brazil
Patients sign consent on their phone

Telehealth

Video consultations happen inside Estação.

Patients get the link on WhatsApp and join from their phone, with nothing to install and no password. You record with one click, and the note comes out ready, just like in person.

01
No app, no password
The link arrives on WhatsApp and opens on the patient's phone.
02
Direct, private call
Peer to peer between the two devices, or through the clinic's own server when the network requires it.
03
The transcript knows who spoke
Your voice and the patient's are recorded separately, with no guessing by the AI.
04
Works with Meet too
In Chrome or Edge, Estação records your microphone along with the meeting tab's audio.

Estação's agents

Good morning
WhatsApp support
Conversational intake
Lab summary
Case study
Sourced copilot
Consultation scribe
Prescription draft
Post-visit summary
Weekly WhatsApp review

Visit the product websiteestacaoclinica.com.br

Featured cases

Retail · Building materials

AstroMatCon

A full ERP for building materials stores, with AI where it saves typing.

Counter sales, inventory, finance, invoicing, WhatsApp and staff time tracking in one system. The AI reads messy input (supplier invoices, job material lists, customer messages) and returns structured data; the server does the math and a person approves.

In production

screens: 179screens
tables: 104tables
automated tests: 1,513automated tests
of invoice items matched without calling AI: 97.5%of invoice items matched without calling AI

View the case

Free-text job material list converted into catalog items, with one item not found and suggestions

Where the AI works

1WhatsApp support bot that answers with the store's real stock, products and delivery info
2Matching supplier invoices to the catalog: rules first, AI only for what's left
3Free-text job material lists turned into a shopping cart
4Nightly catalog sweep: invalid NCM codes, prices below cost, duplicates

Occupational health · Auditing

Sentinela

Nine AI agents audit occupational health records before errors turn into liabilities.

It receives the provider's document package (ASO fitness certificate, clinical form, tests, reports), identifies each page, extracts the fields, checks them against the exam order and the medical protocol, and follows up on missing items with deadlines. Clinical decisions always rest with the doctor.

Being rolled out

specialized agents: 9specialized agents
acceptance criteria: 13acceptance criteria
delivery waves: 5delivery waves
commits: 528commits

View the case

Sentinela9

Maestro

Decision: reads the whole case and decides where it goes next.

Porteiro
Escrivão
Clínico
Auditor
Perito
Elo
Vigia
Analista

Where the AI works

1A squad of agents with defined roles, coordinated by a conductor
2PDFs read by vision models, with a swappable provider per agent
3Quality measured against an answer key before any model change
4Records encrypted at rest

SMBs · Marketing, sales and customer service

Jundo

An AI suite for the whole sales cycle: attract, convert and retain.

Marketing, sales and customer service on a shared core. A team of agents analyzes the brand, media and competitors; the pipeline only moves forward with evidence; support answers on its own and hands off to a person when needed.

In production, launching in October 2026

screens: 81screens
database migrations: 198database migrations
test cases: 12,373test cases
commits: 2,312commits

View the case

Playground with four coordinated AI agents answering a question about campaigns

Where the AI works

1Orchestrated team of agents (Maestro, Mercado, Criativo, Mídia)
2Routing across models from different providers
3Closed decisions with Jev (TypeSafe AI), in shadow mode: in a test with 123 hand-written cases, 8× faster and 70× cheaper
4Lead qualification and automated support

Requirements gathering

Ideia Estruturada

The AI-led requirements gathering that opens every one of our projects.

Before any quote, the AI runs a layered interview about the client's idea: user profiles, modules, flows, rules, integrations and data. The result is a structured technical scope with a complexity score, reviewed and approved section by section. That way the proposal starts from what the client needs, not from assumptions.

In production

layers of questions: 7layers of questions
complexity score: 1 to 10complexity score
screens: 21screens
test files: 22test files

View the case

Ideia Estruturada

01Overview
02Profiles and modules
03Flows and screens
04Rules and behaviors
05Integrations and non-functional requirements
06Entities and data
07Final review

Where the AI works

1Interview led by the AI, which digs deeper into each answer
2Model chosen per layer, across different providers
3Non-functional requirements suggested from ready-made templates: performance, security, LGPD and accessibility
4Validation that flags gaps before the scope is closed

Other systems

Retail network · Tax and BI