Wise.Est Hub

A single-operator operations board. One person, one screen, outcomes only.

Every day it answers four questions:

1. What happened?
2. What should I do next?
3. What is blocking growth?
4. Are we moving toward adoption?

No APIs, queues, or infrastructure surface to the operator. You see the board.

---

Run it locally

You need two terminals.

Terminal 1 — the backend (the brain)

cd backend
python3 -m venv .venv
. .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

The backend is now at http://localhost:8000. It creates its own SQLite file at backend/data/hub.db on first run. The four agents start running on a timer.

Terminal 2 — the frontend (the board)

cd frontend
npm install
npm run dev

Open http://localhost:3000. That is your board.

---

Run it with Docker (one command)

docker compose up --build

Then open http://localhost:3000. The database persists in a Docker volume (hub_data), so your board survives restarts.

---

What the four agents do

Agent	Job	Real?
Scout	Finds real conversations to engage	Yes — live Hacker News, GitHub, Reddit (public APIs, no keys). X is excluded because it needs a paid credential.
Content	Drafts posts from real internal events	Yes — deterministic generator. Drafts always wait for your approval; nothing auto-publishes.
Outreach	Drafts personalized messages for strong opportunities	Yes — deterministic generator. Nothing auto-sends.
Proof Gap	Tracks claims that lack evidence	Yes — seeds the build's load-bearing claims and mines new claims out of your drafts.

The honesty line

This hub will never invent a number or a finding.

• Opportunities are real rows from real public APIs.
• Growth metrics marked derived are counted from real hub activity.
• Growth metrics marked manual (followers, downloads...) are blank until you type in the real number from your own dashboard.
• A claim stays UNPROVEN until you attach real evidence on the Proof Gaps page.

See docs/CLAIM_STATUS.md for exactly what is proven and what is not.

---

Run it 24/7 and open it from your phone

The board can run as an always-on background service on a Mac (auto-starts on boot, restarts itself if it crashes) and be reached privately from your phone over Tailscale — no public hosting, no monthly cost.

1. Install Tailscale on the Mac and the phone, signed into the same account.
2. Build the frontend pointing at the Mac's Tailscale IP: echo "NEXT_PUBLIC_API_URL=http://<YOUR-TAILSCALE-IP>:8000" > frontend/.env.local && (cd frontend && npm run build)
3. Install two macOS LaunchAgents in ~/Library/LaunchAgents — one running uvicorn app.main:app --host <YOUR-TAILSCALE-IP> --port 8000, one running next start -H <YOUR-TAILSCALE-IP> -p 3000 — both with RunAtLoad and KeepAlive true. Load with launchctl bootstrap gui/$(id -u) <plist>.
4. On the phone (Tailscale connected), open http://<YOUR-TAILSCALE-IP>:3000 in Safari → Share → Add to Home Screen. It installs as an app icon named "Wise.Est Hub" and opens full-screen.

Binding to the Tailscale IP (not 0.0.0.0) keeps the board private to your own devices — it is never exposed to public Wi-Fi or the internet. There is no login, so do not bind it to 0.0.0.0 or put it behind a public URL without adding auth.

Tests

cd backend
. .venv/bin/activate
python -m pytest -q

17 tests cover the API, the approval workflow, agent output, proof-gap creation, scoring, and dedup.