The Brainz Portfolio · One roof, many ventures
We don’t extract IP from the operators who built it.
The portfolio · not pre-launch
Real products with real revenue, an AI platform underneath them, and a monitoring stack beside them — each run by the operator who built it, funded by client work, never by outside capital.
29
Ventures across products, platform & services
7
Live today, several already with revenue
$10M COP
Cumulative Lumina sales over ~2 years, ~5× on capital
1×
Roof. Every operator covered at market rate
I
The ventures
The portfolio is the long-term answer to the Basecamp question: we don’t extract IP from the operators who built it. They run it — we back them. Customer-facing products sit on top of a shared AI platform and a self-hostable observability stack. Stage labels are honest: live ships today, building is in active development, frozen waits behind a sequencing gate, parked is captured but not active.
Customer-facing applications, each operated by the person who built it.
Virality scoring + design generation for content teams.
Submit a brief, get three scored design directions back — reel, carousel, static, email, banner — and publish the highest scorer. Built for content teams and marketing directors who want fast, measurable iteration.
Case management + AI legal research for law firms.
A workflow OS for Colombian and LatAm law firms: case management, time tracking, and AI-assisted research on web, mobile, and as a Word add-in. Priced per lawyer, sold as firm packages.
Artisanal fragrance house — and our live dogfood case.
A direct-to-consumer fragrance brand built in Bogotá, sold through Instagram, marketplaces, and retail. It doubles as the proving ground for the Brainz product stack on a real, profitable B2C brand.
Administration software for Colombian buildings and HOAs.
A web and mobile app for property administrators and building associations (juntas) to run residential properties and common areas — the second vertical the distribution machine is pointed at.
An AI finance OS where the agent acts, not just reports.
Turns your plans into queryable, executable financial state — tracking transactions by provenance, multi-currency, with an agent that takes actions via tool use. In daily single-operator use today.
Editorial media on how a Latin team builds with AI.
An independent media property — not a sales channel — documenting how the team builds, ships, and learns in public from Bogotá, with real numbers and editorial independence.
An AI-suggested, AI-built product — the first pattern test.
A product designed and built end-to-end by AI as the first case study for the Brainz idea-factory pattern: AI proposes, AI builds, Brainz validates through outbound.
The engine that finds demand and installs the method.
A 12-week engagement that installs working automation.
Packages the Brainz Method into a consulting offer for LatAm mid-market: AI-powered discovery, framework installation, and hands-on execution that leaves working automation behind.
An AI seller that runs outreach, qualifies, and books meetings.
Automates the first 95% of the B2B funnel — find, enrich, qualify, and book live meetings — then hands showed-up calls to human closers. Distributes for multiple vertical SKUs.
An offer factory for high-ticket, vertical-specific deals.
Produces custom-fitted offers routed from detected signals — AI generates ~80% of the assets, distribution targets the right segments, and the human pod closes only when a fish bites.
The AI orchestration layer that runs beside operators.
Houses the AI agents that support human operators (never above them) and powers internal operations on the Synapse substrate.
Lead enrichment + prospect intelligence for Colombia.
Finds Colombian companies and enriches them with firmographic, financial, and contact data, then scores and segments them — feeding clean, current lists to the AI sales layer. Live at sondea.co.
A large-enterprise digital-transformation engagement.
An enterprise engagement to deliver digital transformation on Brainz AI orchestration — upside, not the core engine.
The shared substrate every Brainz service runs on.
The commercial umbrella — a monorepo of sellable services.
The product/services holding around a Rails monorepo of independently-sellable services, shared infrastructure, and the brand of record.
Unified auth, billing, and API keys for every service.
The foundational layer providing identity, SSO/SAML, API-key management, and billing — one set of credentials across the entire product suite.
Multi-agent orchestration for AI executives.
A Rails-native command center where AI executives live as first-class agents, with organization → project → goal → task workflows and real-time agent conversations.
Encrypted secrets substrate for the platform.
An AES-256 encrypted secrets manager with RBAC, per-project keys, and append-only audit logs — so every service can share a substrate without leaking credentials.
BPMN orchestration + automation flows in one box.
A Rails-native workflow engine that routes processes through BPMN and DAG automation — visual modeler, 630+ connectors, form builder — Camunda + Zapier in one.
AI documentation that stays alive with your code.
Living wikis that auto-generate and update from your Git repo — architecture overviews, diagrams, and a searchable, chat-with-your-codebase interface.
The monitoring stack — LatAm-priced, self-hostable, MCP-first.
Alerting and on-call, self-hosted.
A PagerDuty-shaped alerting brain that evaluates rules against telemetry and fans out across seven channels with escalation and on-call schedules.
Multi-region uptime monitoring + status pages.
Monitors HTTP, SSL, DNS, and TCP from global nodes, opens incidents on failures, and publishes public status pages.
Host and container monitoring for self-hosted infra.
A ~10MB agent that streams CPU, memory, disk, network, process, and Docker metrics every 10 seconds — observability for operators running their own VMs.
Visual regression + AI browser automation.
Captures screenshots, diffs against baselines, and runs autonomous agents in real browsers — bridging the long tail of API-less legacy systems.
APM with real-time traces and metrics.
Application performance monitoring on TimescaleDB — request/job traces, span waterfalls, throughput, error rates, and Apdex — with an MCP server for agents.
Structured, queryable logs with an AI-native query language.
JSONB logs with full-text indexing, session grouping, and a custom query language — exposed via MCP so agents can diagnose production directly.
Error tracking with instant triage.
Fingerprints errors into groups, captures full context, and supports real-time status — a lower-cost alternative to Sentry, with MCP triage tools.
Events, metrics, and anomaly detection.
Track custom business events and build metric dashboards with built-in anomaly detection — gauges, counters, distributions, and alerts when things spike or drop.
Background-job monitoring across Ruby backends.
One dashboard for Sidekiq, Solid Queue, Resque, GoodJob and more — queue size, wait time, duration, failures, and worker health, with MCP retry tools.
Feature flags and gradual rollout for Rails.
A decision layer that answers whether to enable a feature for a user, environment, or segment — A/B tests, progressive rollouts, and targeting.
We don’t extract IP from the operators who built it. They run it. We back them.
The operating principle · Brainz LLC
II
How this works
Active client work — advisory engagements, contractor turnarounds, the BrainzLab platform — covers the team floor at market rate. Surplus routes into the portfolio: the next prototype, domain registration, the first ten customers.
The floor
Client work covers every operator at market rate. We do not arbitrage operator hours.
The surplus
Everything above the floor routes into the portfolio — next prototype, domain registration, first ten customers.
No capital
We do not take outside capital. The portfolio funds itself from the work the team already does.
The other engine
The portfolio is what the operators run. Advisory is the install — the same loop that funds the next product, now on your revenue team.
