Customer completed onboarding.responsepro.app and paid in Stripe

They pick Core or Pro, accept TCPA language, submit the form, then get redirected to Payment Links ($500 setup + monthly). You get the admin email (Resend) and a row in Admin → Onboarding. Use their suggested client_id from the form email, or agree on a final ID before seeding.

Tip: Pricing on the public site is Core $397/mo and Pro $747/mo (plus setup). Match the tier in the seed command to what they paid.

Run seed-client.js to create their dashboard login

From api-backend (Render shell or local with DATABASE_URL):

Args: client_id, pin (their dashboard PIN), tier = core or pro, role = client, IANA timezone, telnyx_number, forward_phone, active. Replace core with pro when needed. Legacy growth still accepted in the script but is deprecated — map old customers to pro.

Timezone (follow-ups + intake SMS wording): choose IANA zone; it updates the seed snippet and SQL below.

If the client already exists from seeding, you can patch numbers only:

telnyx_number = the Telnyx DID (calls + SMS). forward_phone = contractor cell for voice bridge in /api/voice/answer.

Quiet hours (automated follow-ups): nothing sends between 9 PM and 8 AM local for that timezone; due sends reschedule to 8 AM local. Intake “business hours” SMS uses a separate weekday window in the Worker (see docs/WORKER-API-PARITY.md).

Verify Admin → Onboarding and Ops → Client directory

Onboarding tab: their submission row. Ops: client_id appears, tier matches what they bought, active is correct. If missing, the seed failed — check DATABASE_URL and script output.

Test their login URL

Incognito: ?client= + &token= triggers POST /api/auth/login with that PIN and stores a JWT. Replace THEIR_PIN with the same PIN you hashed in seed-client.js:

Core vs Pro UI: the Missed tab is hidden for Core (missed-call SMS is Pro-only in code). They still see Overview, Leads, reports, etc.

Give them the production intake URL

Production hostname: https://intake.responsepro.app/CLIENT_ID (POST). Workers dev URL for debugging: https://responsepro-intake.bluediamondcart.workers.dev/CLIENT_ID. Path segment must match clients.client_id exactly.

Point the website / Angi / Zapier webhook at that URL

POST JSON or application/x-www-form-urlencoded. Common fields: name, phone (required for SMS), email, service, source, city — the Worker normalizes aliases (see restless-tooth-94b4/src/index.js).

Cloudflare Worker secrets (Render parity)

Workers → responsepro-intake → Settings: plain var API_BASE = https://responsepro-api.onrender.com. Secret API_SECRET = same value as Render’s API_SECRET (Bearer for internal routes). Outbound SMS is not sent from the Worker anymore — it goes through POST /api/internal/outbound-sms on Render (caps + compliance). Remove TELNYX_API_KEY / TELNYX_MESSAGING_PROFILE_ID from the Worker if still present.

Full contract: docs/WORKER-API-PARITY.md.

Test intake end-to-end

POST a test lead (real phone). Confirm: SMS to lead + optional owner alert, a lead row on the Leads tab, and a row in follow_up_queue. Check Error Log if anything fails.

Confirm intake is live

Worker returns 200 {"received":true} immediately and finishes work in the background (dedup, opt-out check, lead-event, SMS via API, follow-ups enqueue). Tier limits: monthly lead/SMS caps apply on the API side.

Set inbound webhook in Telnyx Messaging Profile

Telnyx Portal → Messaging → Messaging Profiles → find their profile → set the Inbound Webhook URL to: https://responsepro-api.onrender.com/api/sms/inbound (POST)

Set inbound webhook in Telnyx for their number

Telnyx Portal → Numbers → their number → Edit → set Messaging Webhook URL to the same URL: https://responsepro-api.onrender.com/api/sms/inbound. Method: POST.

Test STOP / START (authoritative suppression)

Text STOP to their Telnyx number. Confirm a row in opted_out_phones. Then text START — row removed, opt_in_events logged, confirmation SMS sent per API.

Test HELP reply

Text HELP from your phone to their Telnyx number. Confirm you receive the help response with support contact info. The server handles STOP/START/HELP automatically — no external scenario needed.

Pro tier only — missed-call recovery SMS

Core accounts do not get automated missed-call texts from /api/voice/event (tier gate in the API). Voice answer still bridges callers to forward_phone for all tiers. For Core clients, skip SMS expectations in tests below unless you upgrade their tier to pro.

Create a Telnyx Call Control Application

Telnyx Portal → Voice → Call Control → Create Application. Set:

Tip: One Call Control Application covers all clients — you don't need a new one per client.

Link their Telnyx number to the Call Control Application

Telnyx Portal → Numbers → find their number → Edit → assign it to the Call Control Application you just created. This tells Telnyx to hit your webhook URL when that number receives a call.

Confirm env vars are set in Render

Render Dashboard → ResponsePro API → Environment → confirm this is set:

• TELNYX_WEBHOOK_PUBLIC_KEY — your Ed25519 public key from Telnyx Portal → API Keys → Webhooks. Required — server rejects all voice webhooks without this.

Tip: One-time setup value — same for all clients. The server handles missed call SMS directly via /api/voice/event (no external scenario needed).

Optional — customize missed-call SMS copy

Production text is hard-coded in api-backend/server.js inside /api/voice/event (mentions client_id, TCPA line, rates). Change it there and redeploy if you want a different template per brand.

Test by calling the Telnyx number and not answering

Pro: Let it ring out (or timeout). Check Missed Calls tab, recovery SMS to the caller, clean Error Log. Core: call still bridges on answer; expect no auto missed-call SMS from the API.

/api/voice/answer returns a connect action to forward_phone (25s timeout) so inbound callers reach the contractor’s cell; no timezone logic on that bridge.

Verify pg_cron job is active in Supabase

Supabase Dashboard → SQL Editor → run: SELECT * FROM cron.job; Confirm there is a job that calls POST /api/internal/follow-ups/process every 5 minutes. This replaces the old Make.com follow-up scenario entirely.

Confirm cadence + tier step cap

Delays between steps (when running): ~1 hour, 24 hours, 72 hours, then 7 days. Pro runs all four steps. Core runs two steps only (API completes the sequence at tier cap). Stops when the lead texts your Telnyx number (logged in lead_replies), on STOP, after the last allowed step, or when monthly SMS caps block sends.

Test follow-up sequence end-to-end

Trigger a test lead via Worker. Confirm follow_up_queue row step = 1, first next_send_at ~1 hour out. After cron fires, step advances until Core cap or Pro completion. Text STOP to verify cancellation.

Orchestration is Supabase pg_cron + pg_net calling POST /api/internal/follow-ups/process with Bearer API_SECRET.

Send the client their dashboard login URL

Tip: Share over a secure channel — token is their raw PIN. Bookmarking logs them in via /api/auth/login.

Walk them through the dashboard (5 min)

Show them: Lead Analytics KPIs, the weekly chart, Missed Calls tab, and how to read response times. Keep it simple — they care about leads contacted and response speed.

Confirm billing in Stripe

They should have completed onboarding.responsepro.app → Stripe Payment Links ($500 setup + $397/mo Core or $747/mo Pro). Match Stripe Customer to their client_id in your records; rotate links in onboarding/script.js if you create new Products/Prices.

Update onboarding status to Active

Onboarding tab → find their row → click Active. This is your internal record that they're fully live.

Schedule 7-day check-in

Set a reminder to follow up in 7 days. Ask: Are they getting SMS alerts? Any issues? This is also your shot to get a testimonial early.

Give your hire their own dashboard login

They only see Prospects, Call Script, Objections, and Pipeline—so they can work leads without touching customer analytics or your admin tools.

Run in PowerShell from your backend folder (same as seeding a client; role is partner in the DB):

One login per rep

Each salesperson should have their own ID and PIN so pipeline ownership and activity stay clear.

Point leads at this rep

In the database, each prospect row belongs to a rep by ID. Assign the territories or lists this hire should work so their Prospects tab fills in.

Confirm they have enough accounts to call

Before go-live, sanity-check counts so the rep isn’t staring at an empty list.

Send them the login link

They sign in with the same dashboard URL you use—only the tabs change based on their role. Replace REP_PIN with their real PIN when you share (use a secure channel).

Watch them click through once

Have them open Prospects, move a stage, skim Call Script and Objections, and check Pipeline. Fix access issues before they’re solo on the phones.

Set expectations

Confirm how you want notes entered, stages used, and follow-ups logged. This CRM is internal—your process, your rules.