The problem
Deciding what to build or pursue next across a portfolio of projects means knowing what you already have, reusable proof and real skills, and weighing it honestly against outside opportunities. That's normally manual archaeology across scattered repos and pasted job listings.
The core layer
Market Scout scans user-selected local project roots into an inventory of proof and skills, then evaluates pasted opportunities for fit, risk, and skill gaps against that inventory. It scores productization candidates and release gaps, and produces weekly trend snapshots. Every agent is deterministic except 1 explicitly gated “Project Reviser” step, which may call a user-owned local model to propose a file revision, never apply one automatically.
The implementation
It's a FastAPI dashboard over a local SQLite inventory and an append-only audit log, with hard safety boundaries: no scraping restricted platforms, no auto-apply, no outbound job applications or outreach, and runtime data like databases and credentials excluded from the repo. It ships its own publish-safety check and a pytest suite to run before sharing anything.
0 likes · 13 views · Aug 25, 2026
An AI-assisted tool for product research and structured strategic next steps.