RAG Platform: AWS Cost Calculator

A retrieval-augmented generation stack has five cost centres: embedding your knowledge base, storing the vectors, retrieving them, generating answers with an LLM, and the plumbing around it. Put in your numbers and see the monthly bill and where it actually goes. Every AWS price is an editable default.

Knowledge base

Documents in knowledge base

Avg tokens per document (~750 = 1 page)

Embedding model

Embedding price ($ / 1M tokens)

Query volume

Queries per month

Retrieved context per query (tokens)

Avg question length (tokens)

Answer length (output tokens)

Generation model (Bedrock)

LLM input price ($ / 1M tokens)

LLM output price ($ / 1M tokens)

Infrastructure · monthly

Vector store

Vector store cost ($ / month)

Object storage & transfer (S3 + egress)

Compute / orchestration (Lambda, Fargate, $ / month)

Monthly cost breakdown

Adjust inputs — full breakdown unlocks when you submit.

Total monthly cost

$0/mo

Cost per query

One-time indexing

Biggest line item

—

Adjust the inputs and submit your email to see the full cost breakdown.

🔒

Submit your email to unlock the cost model

Get the full cost model

We'll email the detailed breakdown and have an AI engineer size your RAG platform against your real corpus and traffic. No pitch.

Work email

No spam. We'll follow up only if it's relevant.

Estimates only — us-east-1 on-demand list prices as starting defaults. Negotiated, provisioned, committed-use, and reserved pricing will differ, and AWS rates change. This is planning guidance, not a quote or an AWS bill. Re-run with your own numbers. Logiciel Solutions · logiciel.io

Link copied