Common issues

"Stream not configured"

Symptom: any stream.* tRPC procedure throws PRECONDITION_FAILED.

Fix: secrets aren't deployed. Run bun run deploy after setting STREAM_API_KEY + STREAM_API_SECRET. See Secrets.

"Going live..." stuck

Symptom: dashboard shows "Going live..." indefinitely. OBS connected, but channel page doesn't flip to LIVE.

Cause: GetStream webhook URL not configured, or pointing at a dead worker.

Fix: GetStream dashboard → Video → Webhook URL → set to https://tv-api.<your-account>.workers.dev/api/webhooks/getstream. Subscribe to call.live_started, call.session_started. See GetStream.

Magic-link emails never arrive

Symptom: invitee clicks "Send sign-in link", nothing in inbox.

Causes (most common first):

1. RESEND_API_KEY not deployed
2. MAIL_FROM domain not verified in Resend
3. Recipient blocked the sender

Fix: check Resend dashboard "Logs" tab for delivery status. See Resend.

ERR_TOO_MANY_REDIRECTS after setup

Symptom: finishing the wizard sends you into a redirect loop.

Cause: session cookie can't cross between web (tv.*) and api (api.tv.*) hosts on *.workers.dev. Public Suffix List blocks cross-subdomain cookies on workers.dev.

Fix: either attach a custom domain (cookies work fine) or use the same-origin proxy in dev (web on :3001 proxies /api to api on :3000).

Migration "no such column"

Symptom: SQL error like no such column: chat_message_count.

Cause: new schema deployed but migrations not applied.

Fix: bun run db:migrate:remote. Check pending: bun run db:migrations:list.

Setup wizard locked

Symptom: /setup returns "already configured", but you want to redo setup (e.g. testing fresh).

Cause: setup_completed_at is set on channel_config.

Fix (destructive — wipes broadcaster):

bun run db:execute:remote --command 'UPDATE channel_config SET setup_completed_at = NULL WHERE id = "site"'

Wizard re-opens. The previous broadcaster's user row + profile is still there. Either work around it or:

bun run db:execute:remote --command 'DELETE FROM user'
bun run db:execute:remote --command 'DELETE FROM channel_config'

(deletes EVERYTHING — only do this in test environments)

CI deploy fails on first run

Symptom: bun run deploy in GitHub Actions fails with "resource already exists".

Cause: alchemy hasn't yet adopted resources you created locally.

Fix: alchemy adopts on adopt: true config (already set in alchemy.run.ts). If still failing, the state file is missing — CloudflareStateStore worker either deleted or not initialized. Push ALCHEMY_STATE_TOKEN secret + re-run.

Worker over CPU limit

Symptom: worker logs show "exceeded CPU time".

Cause: typically a single-request hot loop, or unbounded query.

Fix: check tRPC procedure performance. Common offenders: panel reorder writing too many rows, emote pipeline timing out on a slow provider, snapshot prune query without an index.

Other issues

• OBS not pushing → OBS not pushing
• Chat not working → Chat not working
• Bug not listed here → Open an issue