Stack recommendation

Cheapest stack for a AI chatbot / RAG app

App + vector DB + storage + inference (hosted LLM API or self-hosted GPU).

MVP Traction Scaling

Cheapest

Recommended

$172/mo

incl. $18/mo egress

VPS / compute
medium · Hetzner (On-demand)
$7/mo
Managed database
db_medium · Supabase (On-demand)
$25/mo
Object storage
obj_free_egress · Wasabi (On-demand)
$1/mo
LLM API
llm_mid · Anthropic (On-demand)
$120/mo

Assumptions (Traction)

vps: hours=730
managed_db: hours=730, engine=vector
object_storage: storageGb=200, egressGb=500
llm_api: mTokensIn=50, mTokensOut=20

Editable usage sliders are coming to /build; today these are sane defaults. Costs are effective monthly estimates, not quotes.