🕉

Deployment

Live URLs

Service	URL
Chat UI	veda-guru-ai-ui.vercel.app
Maharshi voice	veda-guru-ai-ui.vercel.app/maharshi
API	veda-guru-ai-api.onrender.com
Docs	veda-guru-ai.github.io/veda-guru-ai-docs

Backend API (Render — free tier)

Repo: github.com/veda-guru-ai/veda-guru-ai-api

Settings used:

Setting	Value
Runtime	Python
Build command	`pip install -r requirements.txt`
Start command	`uvicorn main:app --host 0.0.0.0 --port $PORT`
Plan	Free

Required environment variables:

Variable	Value
`SARVAM_API_KEY`	(your Sarvam AI API token)
`SARVAM_API_URL`	`https://api.sarvam.ai/v1/chat/completions`
`FRONTEND_URL`	`https://veda-guru-ai-ui.vercel.app`

Optional environment variables:

Variable	Purpose
`TAVILY_API_KEY`	Primary web search (free tier: 1000 searches/mo)
`JWT_SECRET`	Auth token signing key (auto-generated if unset)
`DATABASE_URL`	Neon/Supabase Postgres connection string (for persistent data)

Database persistence

By default, the app uses SQLite (veda_guru.db) which is ephemeral — data is lost on restart. For persistent storage, set DATABASE_URL to a Neon Postgres connection string:

postgresql://user:password@ep-xxx.region.aws.neon.tech/dbname?sslmode=require

Neon free tier: 500MB storage, no credit card required. See neon.tech.

Frontend UI (Vercel — free tier)

Repo: github.com/veda-guru-ai/veda-guru-ai-ui

Connect repo to Vercel → it auto-detects static files
Deploy — no build command needed
Update API_URL in script.js and maharshi.js if the backend URL changes

The frontend includes:

Landing page (index.html — served at /, dharmic-themed entry)
Chat page (chat.html — served at /chat, main Q&A interface)
Gurukula learning page (learning.html — served at /learn, learning paths)
Maharshi voice page (maharshi.html — served at /maharshi, voice chat)

Routing via vercel.json:

{
  "rewrites": [
    { "source": "/chat", "destination": "/chat.html" },
    { "source": "/learn", "destination": "/learning.html" },
    { "source": "/maharshi", "destination": "/maharshi.html" },
    { "source": "/(.*)", "destination": "/index.html" }
  ]
}

Documentation (GitHub Pages)

Repo: github.com/veda-guru-ai/veda-guru-ai-docs

GitHub Settings → Pages → Source: GitHub Actions
Add the Jekyll workflow (GitHub suggests it automatically)
Pushes to main auto-deploy the docs site

Redeploying after changes

All three services auto-deploy when you push to the main branch of their respective repos. No manual steps needed.

Render cold start

The free tier spins down after 15 minutes of inactivity. The first request after idle triggers a cold start (30-60s). The frontend handles this with a retry loop and status indicator.