Bennettxai/FounderOS-DEMO

An open-source, single-operator business command center: run a one-person operated company as AI-assisted departments (comms, funnel, social, finances, agents, and a knowledge graph) from one live dashboard.

Founder OS – a demo of an AI‑augmented personal operating system

What it is – Founder OS is a Next.js web app that pretends to run a solo‑entrepreneur’s whole business from a single screen. The UI is split into “departments” (comm‑s, sales funnel, social growth, finances, knowledge graph, etc.) and each department is backed by a small LLM‑driven agent that can run code, query a knowledge base, and update a structured memory layer.

Why it matters – It shows how a collection of connectors (email, Slack, Stripe, Notion, CRM, social APIs) and a hybrid knowledge system (markdown‑based G‑Brain + a governed “Optimal Engine” memory) can be wired together so that autonomous agents have up‑to‑date data and a trustworthy memory. The repo is a fully runnable demo: seed data is loaded into a local SQLite DB, so you can explore every page without any external keys.

Core concepts

Concept	Role
Agents	Small TypeScript classes with a `run()` method that call the Vercel AI SDK (LLM) and interact with connectors and the knowledge layer.
Connectors	20+ integration modules (email/IMAP, Slack, Stripe, Notion, calendar, CRM, social). Each returns a typed `ConnectorStatus` (`connected`, `not_configured`, `error`).
G‑Brain	Markdown files are chunked, embedded, and stored in a vector store. It serves both keyword and semantic search; falls back to local grep if the vector service is down.
Optimal Engine	A “governed memory” that promotes raw signals → claims → verified facts → durable memories. Agents can write signals/claims, but only reviewed facts become part of the OS’s trusted state.
Repository layer	All DB access goes through typed repositories (`agents`, `departments`, `social`, etc.) so swapping the seeded SQLite for a production database requires only a change in `lib/data.ts`.

Main UI routes (demo)

/ – Operator console (pulse, connections, agent roster, knowledge core)
/comms – Unified inbox aggregating email, Slack, WhatsApp, dictation
/funnel – Visual client‑journey flow (linear and radial views)
/social – Follower charts, audience share, posting cadence
/content – Content pipeline & calendar
/finances – Income/expense charts and category breakdowns
/agents – List of AI agents with live run() state
/tasks – Task board populated by agent output
/brain – Interactive knowledge‑graph view
/workflows – Multi‑step tool workflows
/integrations – Status board for all connectors
plus several admin/roadmap pages.

Tech stack (as listed in the repo)

Framework: Next.js 14 (App Router) with server components
Language: TypeScript
Styling: Tailwind CSS (five built‑in themes, default “Monolith”)
Database: better-sqlite3 for the demo; production expects any managed DB
Validation: Zod schemas on every DB/API boundary
LLM calls: Vercel AI SDK
Graphs: d3‑force for the knowledge‑graph visualisation
Testing: Vitest + in‑memory SQLite
Deployment target (demo): Railway (or any Node host) – production adds separate services for G‑Brain and Optimal Engine.

Getting started (quick‑start from the README)

# prerequisites: Node 18+
npm install
cp .env.example .env.local   # optional – only needed for real connectors
npm run dev   # http://localhost:4100

The first run creates a seeded SQLite DB, so every page shows realistic placeholder data without any API keys. Commands for production build, testing, type‑checking and reseeding are also provided (npm run build && npm start, npm test, npm run typecheck, npm run seed).

Intended use cases

Solo founders who want a single dashboard that automates routine business tasks via AI agents.
Developers looking for a reference implementation of:
- a plug‑and‑play connector architecture with honest status reporting,
- a hybrid markdown‑vector knowledge store,
- a governed memory model (Optimal Engine) that separates raw signals from verified facts.
Educators / bootcamps that teach building AI‑augmented SaaS products – the repo is the “reference implementation” used in the Founder OS cohort.

License

MIT – see LICENSE.

Bottom line – Founder OS is a concrete, open‑source demo of an AI‑driven personal business operating system. It bundles a Next.js UI, a set of LLM‑backed agents, a rich connector library, and a two‑layer knowledge/memory system, all runnable locally with seeded data. It is a solid starting point for anyone wanting to experiment with AI‑augmented workflows for solo enterprises.