Pay As You Go

Simple, transparent pricing

No subscriptions, no minimum commitments. You only pay for what you use. Every account gets the same flat rates across all services.

Document Parsing

Extract structured markdown from PDFs or ordered image batches

Standard
$0.005/ page
Max
$0.008/ page
Experimental
$0.005/ page
Legacy
$0.005/ page
Cost-Efficient
$0.002/ page

Document Indexing

Build retrieval indexes on parsed documents

Vector
$0.10/ 1M tokens
Hybrid
$0.50/ 1M tokens

Search

Query your indexed documents

Vector
$0.01/ query
Hybrid
$0.05/ query

AI Responses

Generate answers grounded in your documents

Fast - Input tokens
$1.00/ 1M tokens
Fast - Output tokens
$5.00/ 1M tokens
Pro - Input tokens
$3.00/ 1M tokens
Pro - Output tokens
$16.00/ 1M tokens

How it works

1

Upload & Parse

Upload a PDF or an ordered image batch and we extract structured markdown. You are charged per page based on the parsing method.

2

Index

Build a retrieval index on your parsed document. Choose Vector for speed or Hybrid for deeper understanding.

3

Search & Ask

Search across your documents or get AI-generated answers grounded in your content. Billed per query and per token.

Enterprise & High Volume

Processing large document volumes or need dedicated infrastructure? We offer custom contracts with volume discounts, priority support, and SLAs tailored to your organization.

Contact Sales

Frequently asked questions

Do I need a subscription?

No. Infratex is entirely pay-as-you-go. Add credit to your account and use it across any service. There are no monthly fees or minimum commitments.

What payment methods do you accept?

Currently, credit is added via promotional codes or admin-issued balances. Direct card payments are coming soon.

What is the difference between Vector and Hybrid?

Vector uses classic vector-based retrieval, which is fast and cost-effective. Hybrid adds structural document understanding with AST-based indexing for deeper, more accurate results on complex documents.

How are response tokens counted?

Input tokens include the retrieved context and your question. Output tokens are the generated answer. Both are metered separately at their respective rates.