Applications — small scale stack.
For each requirement below, pick the option that fits your build — recommended first, then free and cheaper alternatives — or skip what your project doesn't need. Tap the info icon next to any requirement to see why it matters.
Web Hosting / Frontend Deployment
Where your code actually runs and serves requests. Picking the right host affects speed, scaling, and how much ops work you do.Pricing & free-tier limitsFree: Vercel Hobby 100GB BW, 6000 exec hrs; Cloudflare Pages unlimited free. Paid: Vercel Pro $20/seat/mo + $40/100GB BW overage, $20/1000 GB-hrs overage. Amplify $0.15/GB served + $0.30/build min. 2026 Pro ~$20-25/mo
Primary Database (Postgres - User Data)
Persistent storage for your app’s data — users, products, orders. The single most important architectural decision for most projects.Pricing & free-tier limitsSupabase Free: 500MB DB, 50k MAU, 2GB egress, 50MB file. Pro: $25/mo incl 8GB DB, 100k MAU, 250GB egress, 100GB file. Overage: $0.125/GB DB, $0.09/GB egress, $0.021/GB file. Neon Free 0.5GB, paid $19/mo 10GB. RDS db.t3.medium ~$60/mo + storage $0.115/GB
Authentication (User Auth + SSO)
How users sign up, log in, and are authorized. Getting roles and access control right early prevents painful rewrites.Pricing & free-tier limitsClerk Free: 10k MAU, 10 orgs. Pro: $25/mo base + $0.02/MAU over 10k, SSO $100/mo add-on B2B. Auth0 Free 25k MAU, B2C Pro $35/mo + $0.07/MAU 1k-10k, Enterprise SSO $0.10/MAU. Auth.js free self-host + DB cost.
File Storage (User Uploads / Assets)
Where you keep files users upload or you serve — images, videos, documents — and how fast they reach visitors around the world.Pricing & free-tier limitsR2 Free: 10GB storage, 10M Class A ops, 10M Class B, $0 egress. Paid: $15/mo base? Actually $0.015/GB-mo storage, $4.50/M Class A, $0.36/M Class B, $0 egress. S3: $0.023/GB-mo storage, $0.09/GB egress, $0.005/1k PUT, $0.0004/1k GET. Supabase Free 1GB, Pro 100GB incl $0.021/GB overage
Vector Database (RAG - User Docs)
Persistent storage for your app’s data — users, products, orders. The single most important architectural decision for most projects.Pricing & free-tier limitsPinecone Free: Starter $0 - 100k vectors (~400MB), 1 index. Serverless: $0 base, min $70/mo effective - $0.08/GB stored, $0.20/1M reads, $2/1M writes. Qdrant Free: 1GB RAM, 100k vectors, 1 cluster. Paid: $28/mo 2GB RAM cluster. Supabase pgvector free within DB. Weaviate Serverless free 1M vectors 14d, paid $25/mo + $0.09/1k queries
LLM API Gateway / Router (Multi-provider failover)
Controls and protects your APIs — quotas, abuse prevention, and firewalls. Important once you have real traffic or many clients.Pricing & free-tier limitsLiteLLM self-host free - only VM cost $10-20/mo. OpenRouter no sub: GPT-4o $2.50/1M in $10/1M out (via OR), Claude 3.5 Sonnet $3/1M in $15/1M out, 5% markup. Portkey Free 10k req/mo, Pro $199/mo 1M req incl + $0.0001/req overage, cache saves 60%. Bedrock $0 per gateway + model fees
LLM Inference Provider - Chat (Core)
Powers AI features — model access, embeddings, and inference. Costs scale with usage, so watch the meter.Pricing & free-tier limits2026 pricing: OpenAI gpt-4o $2.50/1M input $10/1M output; gpt-4o-mini $0.15/1M in $0.60/1M out; o1 $15/1M in $60/1M out. Anthropic Claude 3.5 Sonnet $3/1M in $15/1M out, Haiku $0.25/1M in $1.25/1M out. Groq free: 14.4k req/day Llama 3.1 8B @ 1000 tok/sec. Together $0.18/1M Llama 8B. $5 trial credits each. Overage = token usage
Inference Caching (Redis Prompt/Response Cache)
Stores frequent results in fast memory so you serve less from the database. Big lever for speed and cost.Pricing & free-tier limitsUpstash Free: 500k commands/mo, 256MB max, 1k req/day limit, 20 conn. Pro: $10/mo 100k cmds/day, 10GB, $0.20/100k cmds overage. Redis Cloud Free 30MB. ElastiCache Serverless: $0.125/GB-hr storage + $0.34/1M ECPUs (Redis). Cache hit saves 80% LLM cost, avg 40% hit rate for AI app
Background Jobs & AI Task Queue (Async generations)
Runs slow work (emails, exports, payouts) in the background so users don’t wait. Keeps your app snappy.Pricing & free-tier limitsInngest Free: 50k steps/mo, 100 concurrency, 10k events. Pro $20/mo incl 250k steps + $0.00008/step overage ($80/1M steps). Trigger.dev Free 10k runs/mo, Pro $29/mo 25k runs + $0.0006/run overage. BullMQ self-host free + Upstash Redis $10/mo
Secrets Management
A piece of your stack you may or may not need, depending on scope. Pick the option that fits — or skip it if your project doesn’t require this capability yet.Pricing & free-tier limitsInfisical Free: 5 projects, unlimited secrets self-host, Cloud Starter $19/mo 15 projects, 3 envs. Doppler Free: 25 secrets, 1 project, Team $12/user/mo unlimited. Vault HCP: $0.03/hr (~$21.90/mo) cluster + $0.05/1k API ops, free tier 25 secrets dev
CI/CD & Git Hosting
Where your code actually runs and serves requests. Picking the right host affects speed, scaling, and how much ops work you do.Pricing & free-tier limitsGitHub Free: 2000 min/mo Linux x64 (500MB cache), 500MB Packages. Team $4/user/mo 3000 mins + $0.008/min Linux overage, $0.08/min macOS. CircleCI free 6000 credits/wk, Pro $15/mo 30k credits. Vercel builds included in Pro plan
Object Storage CDN / Asset Delivery
Where you keep files users upload or you serve — images, videos, documents — and how fast they reach visitors around the world.Pricing & free-tier limitsCloudflare Free CDN: unlimited BW, 5 WAF rules. R2 Free 10GB storage + 10M ops. Paid: R2 $0.015/GB-mo + $4.50/1M Class A. CloudFront: $0.085/GB first 10TB + $0.01/10k req + $0.02/10k HTTPS. Savings 100% egress vs S3 for AI images
Product Analytics (Behavior + LLM Usage)
Understands your users and supports them. Drives product decisions and retention.Pricing & free-tier limitsPostHog Free: 1M events/mo, 5k session replays, 1M feature flag req. Paid: $0 + $0.000248/event after 1M ($248/1M events volume down to $0.000075). Self-host free + $20 VM. Mixpanel Free 1M events, Growth from $28/mo. Plausible $9/mo 10k pageviews
Primary Monitoring & Error Tracking
Tells you when things break and why. You cannot fix what you cannot see — this is how you keep downtime short.Pricing & free-tier limitsSentry Free: 5k errors/mo, 10k replays, 1M spans. Team $26/mo 50k errors incl + $0.00029/error overage, 100k replays + $0.00036/replay. Business $80/mo 500k errors. GlitchTip self-host free ~$10/mo VPS. Datadog APM $31/host/mo + $1.70/1M spans
LLM Observability (Langfuse / Helicone)
Tells you when things break and why. You cannot fix what you cannot see — this is how you keep downtime short.Pricing & free-tier limitsLangfuse OSS free self-host ~$20/mo VM. Cloud Hobby Free: 50k traces/mo, 1k evals. Pro: $39/mo 500k traces, 10k evals, $0.00006/trace overage ($60/1M). Helicone Free 100k req/mo, Pro $20/mo 500k + $0.00005/req. LangSmith Plus $39/mo 10k traces $0.50/1k overage. Critical for token cost tracking
Search (In-App & Docs Search)
Lets users find things fast. A slow or missing search directly costs engagement and conversions.Pricing & free-tier limitsTypesense Cloud Free: 1 cluster 0.5GB RAM, 5k docs. $19/mo 2GB RAM 100k docs + $0.20/100k searches overage. Meilisearch OSS free self-host ~$10/mo VM. Algolia Free 10k search req/mo 10k records, Build $0.50/1k req + $0.40/1k records/mo after, Growth $150/mo incl 100k req
Email API (Transactional, Invites)
Sending transactional email (receipts, resets, alerts) and sometimes marketing mail. Deliverability matters more than features here.Pricing & free-tier limitsResend Free: 3k/mo, 1 domain, 1 day log. Pro $20/mo 50k incl + $0.0004/email overage ($0.40/1k). SES: $0.10/1k emails + $0.12/GB attachments, free tier 3k/mo if EC2. SendGrid Free 100/day, Essentials $19.95/mo 50k + $0.45/1k over. Postmark $15/mo 10k
Billing / Usage-based Metering (AI Credits)
Powers AI features — model access, embeddings, and inference. Costs scale with usage, so watch the meter.Pricing & free-tier limitsStripe Billing: 0.5% of recurring volume (capped), usage-based metering free: 1000s of meter events free. Polar free: 4% + Stripe fees (0.7% + 30c). LemonSqueezy 5% + 50c. In test: free unlimited. Paid: Metered billing $0 cost but Stripe fees 2.9% + 30c per charge. Metronome Enterprise $2000/mo+ for complex AI credits
API Gateway + Rate Limiting (Per-user AI quotas)
Controls and protects your APIs — quotas, abuse prevention, and firewalls. Important once you have real traffic or many clients.Pricing & free-tier limitsCloudflare Workers Free: 100k req/day, 10ms CPU. Paid: $5/mo 10M req incl + $0.30/1M overage, $0.15/1M rate limit rules evaluations. Kong OSS free, Cloud $6/gateway/mo + $250/1M API calls. AWS API GW HTTP: $1/1M req + $3.50/1M REST, $0.09/GB. Critical to prevent LLM cost abuse: per-user 100 req/min
How this small scale applications checklist works.
Each requirement below is something a small scale applications build typically needs. Pick one of the four researched options — recommended, free, cheaper or paid — add your own with "Other", or skip the requirement if your project doesn't need it. Nothing is mandatory; the plan on the right tracks what you've decided so nothing gets forgotten.
Your picks are saved in this browser automatically, so you can come back anytime. Options are researched per build level and refreshed as vendors change their plans — always verify details on the provider's page before committing.