# Friendly Experts > Friendly Experts fixes apps built with AI tools (Lovable, Replit, Bolt, v0, Cursor, Windsurf, Claude Code, Base44) at fixed prices, and sets up AI for small businesses. Start with a free, read-only health check: https://friendlyexperts.ai/health-check/ ## How it works - Free Health Check ($0, report within 24 hours): Read-only scan of your repo link or live URL. Exposed keys, Supabase RLS per table, build and deploy status. Known-vulnerable packages, speed and accessibility score, broken links and console errors. One-page red, yellow, green report with your top 3 problems and the fixed price for each. No code changes, no call needed. - Pick a fix from the menu. One fix covers one root cause. One fix means one root cause. If it turns out to be three problems, we stop and quote before touching anything else. - Guarantee on menu fixes: Fixed or it's free. On any menu fix, if the problem isn't fixed inside the stated window (with repo access and a way to reproduce it), you get a full refund. On a Full Rescue, if any P0 is still open at handoff, the second half is waived. ## Price menu (USD) - Full Audit: $249, 48 hours. 100% credited to any fix or rescue booked within 14 days. - Fix, Deploy, build or domain: $199, 48 hours. Failing builds, env vars, Vercel, Netlify and Replit deploys, custom domain and SSL, export off the builder to your own GitHub. - Fix, Login and auth: $299, 48 hours. Supabase, Clerk and Firebase auth, OAuth and Google sign-in, magic links, session loops, password reset, protected routes. - Fix, Database and Supabase: $299, 48 hours. RLS policies, schema fixes, broken queries, connection limits, data that saves but never shows. Includes RLS tests so it stays fixed. - Fix, Payments and Stripe: $399, 48 hours. Checkout, subscriptions, signed and idempotent webhooks, test mode to live mode, post-payment access, failed-payment emails. - Fix, Speed and performance: $299, 48 hours. Slow pages, heavy images, slow queries and missing indexes, caching, bundle size. Before and after Lighthouse scores in the report. - Fix, Security lockdown: $449, 72 hours. Turn on and test RLS, move secrets server side and rotate exposed keys, add headers, rate limits and webhook signing. - Rush: +50% of the fix price, 24 hours, start within 4 business hours. - Launch Ready Pack: $999, 5 business days. Full Audit. Up to 3 fixes from the menu. 14 days of follow-up fixes on the same issues. - Full Rescue: $2,999 (half to start, half on delivery), 2 weeks. Full Audit, with every P0 and P1 fixed. Tests on login, payments and your main user flow. CI so future changes get checked. Error monitoring and verified backups. Handoff doc and 30 days of follow-up on anything we touched. - Platform Move: From $5,999, 4 weeks. Move off Lovable, Bolt, Base44 or v0. Onto your own Next.js and Supabase on Vercel. All code in your GitHub. - Care plan, Tune-Up: $149/month. Uptime and error monitoring. Weekly dependency and security patches. Monthly health report and backup check. 20% off any fix. - Care plan, Pit Crew: $449/month. Everything in Tune-Up. Up to 2 menu fixes a month. A monthly 30-minute check-in. ## AI setup for small businesses (USD) - AI Quick Start (Solo owners): $497, 1 week. Pick the right tool (ChatGPT, Claude or Copilot) and plan. Account and privacy settings. A business profile so the AI knows your company. 1 custom GPT or Claude Project for your most common task. A 20-prompt library for your work. One 60-minute live training, with recording and cheat sheet. - AI Team Setup (Teams up to 10): $1,997, 2 weeks. Business workspace with roles and permissions. A one-page AI use policy. 2 custom GPTs or agents. 2 automations (Zapier, Make or n8n) with a human-approve step. 90-minute team training and a written playbook. 30 days of questions. - AI Automation Build (Growing businesses): $4,997, 30 days. 3 workflows built and tested on your real work. Lead follow-up, quoting, inbox triage, reporting and similar. Connected to up to 5 tools, with AI drafting and a person approving. Training, handoff doc and 30 days of support. - AI Help Desk: $499/month, Next business day. Up to 2 small changes a month. Monthly 30-minute tune-up call. New-hire onboarding. A note when your tools change. - AI Help Desk Priority: $1,250/month, Same business day. Up to 6 changes a month. One new automation a quarter. Monthly leadership review. - Guarantee: Working in 30 days or we keep working free. If the agreed workflows aren't in daily use by day 30, we keep going at no charge until they are. ## Fix guides - [Lovable app won't deploy or publish](https://friendlyexperts.ai/problems/lovable-app-wont-deploy/): Most Lovable deploy problems come from four places: changes that were never re-published, custom domain DNS that hasn't verified, environment variables missing on a new host, and single-page-app routes that 404 on refresh. Each has a short, specific fix. - [Lovable + Supabase: "new row violates row-level security policy"](https://friendlyexperts.ai/problems/lovable-supabase-row-level-security-errors/): Row Level Security (RLS) is on, but there is no policy that allows the action. Supabase blocks inserts with "new row violates row-level security policy" and quietly returns zero rows for reads. Add a policy for each action (select, insert, update, delete) that matches who should be allowed. - [Replit app login broken after deploy](https://friendlyexperts.ai/problems/replit-auth-broken-after-deploy/): The deployed app runs on a different domain than the editor preview, so OAuth callback URLs, cookie settings and secrets that worked in development often don't carry over. Register the deployed URL with your login provider, confirm the secrets in the deployment, and set cookies to work behind Replit's proxy. - [Bolt.new Supabase errors: invalid API key, failed to fetch, missing tables](https://friendlyexperts.ai/problems/bolt-supabase-errors/): When a Bolt app loses its connection to Supabase, the cause is usually missing VITE_ environment variables on the host, a paused free project, migrations that never ran, or RLS blocking the query. Check them in that order. - [Taking a v0 app to production on Vercel](https://friendlyexperts.ai/problems/v0-app-to-production/): A v0 app is a Next.js project. To run it in production you need a clean next build, environment variables set in Vercel, secrets kept on the server, image domains allowed, and a real database and login behind any screens that still use sample data. - [Stripe webhook failing: "No signatures found matching the expected signature"](https://friendlyexperts.ai/problems/cursor-stripe-webhook-failing/): Stripe signs the exact bytes it sends. If your framework parses the JSON before you verify, or you use the wrong signing secret, verification fails with "No signatures found matching the expected signature for payload". Verify against the raw body with the secret for that exact endpoint and mode. - [Google sign-in "Error 400: redirect_uri_mismatch" or login goes to localhost](https://friendlyexperts.ai/problems/google-login-redirect-uri-mismatch/): Google only redirects to callback URLs you registered, character for character. If you use Supabase Auth, Google redirects to Supabase first, and Supabase then sends the user to your Site URL or an allowed redirect URL. Both lists have to include your live domain. - [Supabase service role or secret key exposed in your app](https://friendlyexperts.ai/problems/supabase-key-exposed-in-frontend/): The service_role (or new secret) key bypasses every Row Level Security rule. If it appears in browser code, anyone can read and change your whole database. Rotate it, move privileged work to a server or Edge Function, and turn on RLS with policies for every table. ## Pages - [Home](https://friendlyexperts.ai/) - [Price menu](https://friendlyexperts.ai/prices/) - [AI setup](https://friendlyexperts.ai/ai-setup/) - [Free health check and request form](https://friendlyexperts.ai/health-check/) - [About](https://friendlyexperts.ai/about/) - [Privacy](https://friendlyexperts.ai/privacy/) - [Terms](https://friendlyexperts.ai/terms/) ## Contact Use the request form at https://friendlyexperts.ai/health-check/ to book. For other questions, email hello@friendlyexperts.ai. Nothing is charged on the website.