Help

Support

What this page is: the six most common post-setup issues and how to fix each one. What to do here: scan the symptom column, find yours, follow the fix. Why it matters: the failure modes below are all infrastructure-level — they look like code bugs but aren't.

Troubleshooting reference

Symptom

Magic-link emails never arrive

Likely cause

Supabase Auth is still on its default email service (heavily rate-limited — 3-4 emails per hour). Or the from-address points at an unverified domain.

Fix

Wire Resend custom SMTP into Supabase Auth (Authentication → SMTP Settings). Use a sender on a verified subdomain. Don't send from the bare apex — Resend rejects it silently and Supabase surfaces 'Error sending email' that looks like a code bug.

Symptom

Login or signup hangs at the redirect step

Likely cause

The production domain isn't in Supabase Auth's allowed redirect URLs.

Fix

Supabase dashboard → Authentication → URL Configuration. Add the production domain + /auth/callback to the allowlist. Wildcards work for Vercel preview deployments — add https://*-your-team.vercel.app/auth/callback if you preview frequently.

Symptom

Vendor name / email / phone shows up blank in the footer

Likely cause

NEXT_PUBLIC_VENDOR_* env vars not set on Vercel. The template is intentionally vendor-neutral by default so it never leaks identity across deployments.

Fix

Vercel dashboard → project → Settings → Environment Variables. Set NEXT_PUBLIC_VENDOR_NAME / NEXT_PUBLIC_VENDOR_EMAIL / NEXT_PUBLIC_VENDOR_PHONE / NEXT_PUBLIC_VENDOR_CALENDLY. Redeploy to pick up the new values. See /configure for the full list.

Symptom

/api/setup/* returns 401 after the deployment is live

Likely cause

Expected behaviour. Setup routes are protected by SETUP_SECRET and run once during initial provisioning. They are not user-facing endpoints.

Fix

Don't call /api/setup/* in production. If you genuinely need to re-run a setup step, rotate SETUP_SECRET first, run the step with the new secret, then rotate again.

Symptom

AI coaching responses fail with 'invalid API key'

Likely cause

The deployment's ANTHROPIC_API_KEY or OPENAI_API_KEY env var is unset, expired, or has insufficient credit.

Fix

Vercel dashboard → project → Settings → Environment Variables. Confirm both keys are present and active. RaiseReady runs on the deployment owner's keys (BYOK), so Corporate AI Solutions cannot top them up on your behalf.

Symptom

Voice coaching session fails to start

Likely cause

ElevenLabs API key unset, agent ID not provisioned, or the deployment's ElevenLabs account is out of credit.

Fix

Check ELEVENLABS_API_KEY on Vercel + agent provisioning status in the setup wizard's audit log. Voice features are optional — if you don't intend to use voice coaching, you can disable it in the platform config without affecting text coaching.

When self-help isn't enough

The owner of each RaiseReady deployment supports their own end users. Corporate AI Solutions supports the deployment owner. If you are an end user trying to reach support for a coaching programme you joined, contact the operator who invited you — their contact details are in the footer of this deployment.

If you are the deployment owner and have hit something not covered above, set NEXT_PUBLIC_VENDOR_EMAIL in your Vercel env so your customers see your address, and reach out to Corporate AI Solutions through your established channel for template-level issues.

Haven't set up yet?

The setup wizard provisions Supabase + GitHub + Vercel + ElevenLabs in a single guided flow.

Start the setup wizard