FounderOS
AI Chief of Staff for Startup Founders
Demo Video
Screenshots
Overview
A streaming AI pipeline and persistent memory system that gives startup founders a context-aware assistant — remembering decisions, pivots, and milestones so recommendations are specific rather than generic.
Technical Details
Built a streaming AI pipeline using ReadableStream.tee() to fork the model response — one branch streams tokens to the user instantly while the other persists the conversation and extracts context records in parallel, so memory saving never blocks the reply.
Designed a memory system that captures decisions, pivots, and milestones from user actions and injects the 10 most relevant records into every system prompt.
Shipped Stripe subscription billing across four tiers with HMAC-verified webhooks and server-side feature gating, on Postgres with row-level security and an isolated admin service client for privileged routes.