DevOS runs as a split frontend/backend deploy today. Two deploy configs are
committed on purpose — both are live and both are required, this is not
duplicate/leftover config:
Config
What it is
Status
vercel.json
Frontend build (client/dist) + proxies /api/* to the Render API
Live — personal-ai-blue.vercel.app
render.yaml
Backend API (Express server)
Live — personal-ai-f2f9.onrender.com
VPS-DEPLOY.md / .github/workflows/deploy-vps.yml
Planned migration to a self-hosted Vultr VPS
Not live yet — future work, do not treat as active
Why both Vercel and Render: Vercel serves the static frontend and rewrites
/api/* requests to the Render-hosted backend (see the rewrites entry in
vercel.json). Render runs the actual Node/Express API and database access.
Neither one is redundant with the other — removing either breaks the app.
server/lib/env.js is host-agnostic: a deploy counts as production when
NODE_ENV=production (set in render.yaml), and its public URL comes from
PUBLIC_URL or, on Render, RENDER_EXTERNAL_URL. Because that one variable
switches the production safety checks on, the server also refuses to start if it
sees it is running on Render (RENDER=true) without NODE_ENV=production.
The only two live URLs are:
Frontend: https://personal-ai-blue.vercel.app
API: https://personal-ai-f2f9.onrender.com
Troubleshooting
Error
Fix
configured: false in status bar
Run npm run dev from the project root, not inside server/
Error 400: redirect_uri_mismatch
Add http://localhost:3001/api/auth/google/callback to Google Console redirect URIs
Error 403: access_denied
Add your Google account as a Test user in OAuth consent screen
Notion API token is invalid
Regenerate the integration token and re-share databases with it
Gemini 429 rate limit
Router auto-falls back to Groq — check logs for [llm] falling back to groq
Slack messages not arriving
Confirm bot was invited to your DM: /invite @your-bot-name in Slack
GitHub PRs not loading
Check GITHUB_OWNER + GITHUB_REPO are set and token has repo scope