Ragextract Document Processing API

Document Processing API Designed to Reduce LLM OCR Parsing Costs up to 90%⁺

Most AI document services require you to parse the whole document before you can search through and extract what you need. This means you're being charged precious credits for pages you won't ever see.

Ragextract uses an alternative approach.

Designed for long documents of 100+ pages, Ragextract use RAG techniques with a robust document ETL pipeline to provide the retrieval functionality before any LLM/VLM OCR parsing takes place. This lets you "Search before Extract" reducing the number of pages that need to go through your LLM by up to 90%!

Get Started For Free

Request a Demo

14 Days Free Trial

GDPR Compliant

30 days Refund Policy

⁺ subject to use case.

Solutions & Use Cases

Optimise Long Document Processing Automation

Ragextracts Works Wherever there are Documents Involved

Cost Savings when Parsing Applications, Claim Forms and Insurance Policies

For Insurance, Underwriters and Brokers

Insurance company X planned to use VLMs to process up to 20,000 third party claims a year but projected costs would have required 3x their annual budget. They just wanted to extract 6-7 pages of claimant data per document but their proposed solution would have seen them billed for an additional 60 pages of legal filler they couldn't even use.

Ragextract provided the search and retrieval solution which allowed Insurance company X to reduce projected LLM costs by identifying only the relevant pages from each document that needed parsing.

Productivity Boost For Handling Long Contracts, Court Documents and Research

For Legal Teams, Lawyers and Consultants

Law firm X works regularly with documents of all sizes and typically the higher end of document tended to exceed way over 300 pages each. Though their initial lightly specced machines to run processing worked largely without issue, the servers and applications suffered when larger documents were imported, especially at seasonal busy times.

Ragextract provided the infrastructure and capability to handle an assortment of legal documents such as contracts, legal regulations and court documents. Drag and drop via the webUI or push through our API, Ragextract automates the storage, splitting and retrieval for your internal workflows or applications.

secs

Time to Validate & Extract From Tenders and Contracts

For Construction Firms, Property Managers and Sales Teams

Property Management Company X uses Ragextract to increase the number of government tenders they can process through their sales pipeline. Ragextract allows them the ability to not only search through a single document but across multiple documents within the tenders and thus can quickly answer complex pre-qualification questions and data extraction specific to their business quickly and efficiently.

Ragextract provided the perfect pipeline designed to speed up document Q&A workflows as well as optimise data extraction. It can take as little as 30 seconds to process tender briefs, questionnaires, appendixes and site plans to find the information you're looking for.

How it works

Just Drop In Your Files and We'll Do The Rest

Search and Retrieve Pages that You Parse Yourself

Dedicated Document ETL Pipeline

Leverage our purpose built ETL pipeline which is designed to handle documents of up to 5000 pages. Drop us the file directly or send the URL, we'll handle the processing and GDPR compliant storage.

Multimodal Embeddings and Retrieval

Ragextract uses state-of-the-art multimodal embeddings so you can search over images, diagrams, tables, charts, presentations and text. We then host and populate internal vector stores which power our retrieval APIs.

Search API to Identify Relevant Pages

Search over hundreds of pages without LLM OCR parsing to save thousands in costs. Ragextract returns pages as separate binary files which you can combine and feed into your LLM of choice to get the data you need.

Ragextract works even better if you have same documents that can come in different layouts!

Insurance Policies

Supplier Invoices

Financial Statements

Reconciliation Spreadsheets

Bills

Deep Research

Court Documents

Tenders & RFPs

Slide Decks

Sales Presentations

SEC Filings

Commercial Property Contracts

Legal Documents

Construction Questionnaires

Academic Papers

Technical Manuals

And much more!

Simple REST API that works with n8n, Make, Zapier and more

Upload documents directly to Ragextract and we'll take care of the storage, splitting, vectorizing and indexing in a matter of seconds.

curl -X POST https://api.subworkflow.ai/v1/vectorize \
  --header 'x-api-key: $RAGEXTRACT_API_KEY' \
  --form "file=insurance_claim_application.pdf"

Search across 100s of pages instantly using natural language and Ragextract will return only the relevant ones (score-based). No more praying you'll hit the right keywords and images also work - it's all contextual!

curl -X POST https://api.subworkflow.ai/v1/search \
  --header 'x-api-key: $RAGEXTRACT_API_KEY' \
  --data '{
    "query": "Find motor vehicle details and registration of all parties involved in claim",
  }'

Finally, pass those matching pages into an VLM/LLM of your choice. You decide how and what gets extracted depending of your use-case. Ragextract keeps your data safe from leaks by serving them as dynamic and expirying links.

curl https://api.openai.com/v1/responses \
  --header "Content-Type: application/json" \
  --header "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-4.1",
    "input": [
      {
        "role": "user",
        "content": [
          { "type": "input_text", "text": "Extract claimant details in the following structure..." },
          {
            "type": "input_image", // Use Ragextract result here!
            "image_url": "https://api.subworkflow.ai/v1/share/dsx_B5bsOBDzsXsqfmLo?token=kCThsH"
          }
        ]
      }
    ]
  }'

RAG Backend Built For Document Extraction

Ragextract Wants To Be Complimentary in Your AI Toolkit

VS Unstructured.io, LlamaIndex, Reductio

Ragextract is a Document RAG service used to optimise long document extraction workflows

Ragextract challenges the popular document processing model of other providers by avoiding the "parse-first-search-later" approach. In fact, we do the opposite! We believe our method is faster, cheaper and less wasteful.

VS Assistants and File Search Tool

Ragextract focuses on large documents which exceed LLM limits

Ragextract is for the long and large documents that most LLMs typically struggle with - scanned Pdfs, technical manuals, image heavy reports etc. Our users typically prefer our more advanced RAG and retrieval setup as they can do more with the output.

VS Chat-with-PDF Apps

Ragextract doesn't bundle LLMs or pre-baked prompts into its service

Not only does this reduce our costs significantly, our belief is that bundling models is unnecessary as AI Developers actually prefer to handle their own prompts. Ragextract is built for developers for more challenge document extraction workloads.

VS n8n, Make.com, Zapier, no-code tools

Ragextract is your large document processing partner to boost your automation

Many low code platforms aren't designed for heavy document / high cpu-consumption tasks. Typically, they exhaust their limits which then crash your instance. Ragextract pairs perfectly with these services to offer the necessary processing power to make business flow.

VS Building Your Own

Ragextract extends your budget and lets you focus on the AI Experience

Document processing infrastructure consumes a large amount of your team's time, budget and reputation if not done correctly. Adopting Ragextract means you can focus on building a great AI experience for your users and reduce development time by weeks.

VS Your Existing Provider?

Still Unsure? Let's Chat!

We'll be happy to discuss what a migration might look like and give you a quote with timelines.

Request a Demo

Pricing

Simple Pricing that Scales as You Grow

All Features. All APIs. No Hidden Fees.

Monthly

Annually

Save

Save 20%

$15/seat/month

$12

/seat/month

Billed $144 annually

Starter

10 concurrent jobs, max 100mb uploads

Usage

Shared worker pool

100mb upload size limit

1g storage

total

30 day data retention

Unlimited number of pages

Unlimited number of retrievals

Organisation

Up to 5 team members

Unlimited workspaces

10 API keys

per workspace

10 concurrent jobs

per workspace

Admin, manager, developer & readonly roles

SSO/OAuth Logins

Support

Community forum

Email support on best effort basis

No phone support or SLAs offered

Get Started For Free

14 day free trial. Card required.

Save 20%

$49/seat/month

$39.20

/seat/month

Billed $470.40 annually

Standard

100 concurrent jobs, max 500mb uploads

Usage

Shared worker pool

500mb upload size limit

10gb storage

total

Unlimited data retention

Unlimited number of pages

Unlimited number of retrievals

Organisation

Up to 50 team members

Unlimited workspaces

100 API keys

per workspace

100 concurrent jobs

per workspace

Admin, manager, developer & readonly roles

SSO/OAuth Logins

Support

Community forum

Email support on best effort basis

No phone support or SLAs offered

Get Started For Free

14 day free trial. Card required.

Dedicated Support

Base Price

Custom

Quote tailored to requirements

Custom

Tailored to Requirements

Usage

Dedicated worker pool

Custom upload size limit

Custom storage

Custom data rentention

Unlimited number of pages

Unlimited number of retrievals

Organisation

Custom quantity of team members

Unlimited workspaces

Unlimited API keys

per workspace

Custom concurrent jobs

per workspace

Admin, manager, developer & readonly roles

SSO/OAuth Logins

Support

Community forum

Email and private Slack channel

Phone support and SLAs available (+fees)

Get a Quote

or email us at sales@subworkflow.ai

14 Days Free Trial

Cancel Anytime

30 days Refund Policy

Migration Support Available

* "unlimited" has a finite value: unlimited duration means a max of 730 days (2 years) and unlimited quantity is a max of 999999 units unless stated otherwise.

** To ensure high performance for enterprise customers and stability for our non-enterprise customers, dedicated worker clusters are provisioned and operated separately from the standard pool.

FAQs

Not Finding What You're Looking For? Contact Us!

Do you have a free plan?

Not yet! Our generous free trial gives all new customers full access for 14 days to evaluate the service. You can also cancel anytime with a single click.

How many documents can I upload per month?

As many as you like... just as long as you stay within the designated storage limit. We don't charge by the number of pages so whether 10k pages or 100k pages, the price stays the same.

Can I cancel my subscription at any time?

Yes of course! We want you to be happy to stay with us so if anything isn't working right, reach out and let's see if we can fix it or implement a new feature for you. Just head over to the billing page and you can manage your subscription from there.

Why is a credit card needed to sign up for the trial?

Though we'd like to offer an easier alternative, credit card registration is one of the few productive ways to protect the service from sign-up spam and fake users. This in-turn ensures better stability for the all users on our platform.

Will I get charged at the end of my trial?

Yes. If you're happy to continue with Ragextract, we'll automatically handle the billing and you don't need to take any further action. If it turns out Ragextract isn't for you, please cancel your subscription before the trial expires.

I'm not sure if Ragextract is for me. Can I get a demo?

Yes! Please reach out to us at sales@subworkflow.ai and we'll get back to you as soon as we can during business hours. Please include details of your organisation, team size and high level brief of what you're looking for as this will help us tailor the demo for you.

Document Processing API Designed to Reduce LLM OCR Parsing Costs up to 90%+

Solutions & Use Cases

How it works

RAG Backend Built For Document Extraction

Pricing

FAQs

Document Processing API Designed to Reduce LLM OCR Parsing Costs up to 90%⁺