Easiest way to build your own RAG pipeline
Fully managed API for processing, querying and chatting with your documents
# Upload documents
curl -X PUT \
     --url /v1/collections/mycollection/upload \
     --form file=@may_report.pdf \
     --form file=@june_report.pdf

# Start chatting
curl -X POST \
     --url /v1/collections/mycollection/chat \
     -d '{"message":"What is MoM change in spend?"}'
Fully managed
Focus on your product
Outpace competitors
Made for developers
Hands off OpEx

Retrieval Augmented Generation (RAG) allows to enhance existing Large Language Models (LLM) with own data without the need to fine tune a model, which is often time consuming and expensive. With documentAI you can build your own RAG pipeline with single API call.

We manage the full pipeline from document processing, chunking, embedding, querying and augmenting an LLM for chat. You do not need to have any prior knowledge or experience with Artificial Intelligence to harness the power of your documents.

Whether you are planning to upload five documents or tens of thousands we take care of the scalling and maitaining the infrastructure so you can focus on building your product.

Process any document
Upload, Download or Crawl

Ingesting your data is a breeze whether you have it saved locally, hosted or want to craw an entire website - we've got you covered.

Currently we support a number formats: PDF, Word, Excel, Text and Web Pages - including article extraction and are always working on expanding our supported formats.

You are always in control of your documents and can always add or remove documents from your collections to suit your needs.

Measure & Evaluate
Track your accuracy

Evaluating RAG pipelines is crucial to ensure high performance when documents, prompts or LLMs change. Rigorous evaluation provides insight into components by measuring faithfulness, answer relevance, context relevancy and context recall. All of this is automated for your convenience.

  • Faithfulness - Measures factual accuracy of the answer using LLMs to extract statements and validate against context.
  • Answer Relevancy - Measures relevance of the answer to the question by using LLM to generate probable questions for the answer and compare to actual question.
  • Context Relevancy - Measures signal-to-noise ratio in retrieved contexts by using LLM to identify required sentences and compare to total.
  • Context Recall - Measures if retriever found all necessary information by checking if LLM can find statements from ground truth answer in retrieved context.
Provenance
Know where the answers come from

Large Language Models suffer from hallucinations where they make up infromation and pass it off as a fact. With documentAI you can always see where the information came from.

Complete transparency - we store your original files so you can always double check the chatbot's statements against the source. No more guessing if the answers are correct, pull up the precise data behind the chatbot's knowledge.

The key benefit is trustworthiness. With documentAI you get explainability - the ability to validate that responses are grounded in your documents, not imaginary facts. Your customers get reliable, verifiable information, strengthening trust and satisfaction.

{
  "message": {
    "id": "da4c1fcd-6981-4ba4-8e6f-deae2b74824f",
    "date": "2023-09-21T19:26:05.148345Z",
    "content": "The iPhone sales for Apple Inc. for the three months ended July 1, 2023, and the nine months ended June 25, 2022, are as follows:\n\n- Three months ended July 1, 2023: $39,669 million\n- Nine months ended June 25, 2022: $156,778 million\n\nThese figures represent the net sales generated specifically from the iPhone product category.",
    "context": [
      {
        "collectionId": "finance",
        "documentId": "cc10f2f8-0895-4914-a6d2-1a50b4f679c8",
        "chunkId": "36964a76-e8b3-4b1b-af7b-54fd18824d0d",
        "content": "Rest of Asia Pacific  5,630    6,150    23,284    23,002  financeTotal net sales ! 81,797   ! 82,959  ! 293,787   ! 304,182  finance        finance(1) Net sales by category:        financeiPhone ! 39,669  ! 40,665  ! 156,778   ! 162,863  financeMac  6,840   7,382    21,743    28,669 financeiPad  5,791    7,224    21,857    22,118  financeWearables, Home and Accessories  8,284    8,084    30,523    31,591  financeServices  21,213    19,604    62,886   58,941  financeTotal net sales ! 81,797   ! 82,959  ! 293,787   ! 304,182  finance finance finance finance financeApple Inc. financeCONDENSED CONSOLIDATED BALANCE SHEETS (Unaudited) finance(In millions, except number of shares which are reflected in thousands and par value) finance finance July 1, finance2023  September 24, finance2022 financeASSETS: financeCurrent assets:    financeCash and cash equivalents ! 28,408  ! 23,646 financeMarketable securities  34,074    24,658  financeAccounts receivable, net  19,549    28,184  financeInventories  7,351    4,946 financeVendor non-trade receivables  19,637    32,748  financeOther current assets  13,640    21,223  financeTotal current assets  122,659    135,405",
        "metadata": {
          "url": "https://www.apple.com/newsroom/pdfs/fy2023-q3/FY23_Q3_Consolidated_Financial_Statements.pdf"
        }
      },
      {
        "collectionId": "finance",
        "documentId": "cc10f2f8-0895-4914-a6d2-1a50b4f679c8",
        "chunkId": "3aa98809-f553-4c75-a83c-23189929311c",
        "content": "Apple Inc. financeCONDENSED CONSOLIDATED STATEMENTS OF OPERATIONS (Unaudited) finance(In millions, except number of shares which are reflected in thousands and per share amounts) finance Three Months Ended  Nine Months Ended finance July 1, finance2023  June 25, finance2022  July 1, finance2023  June 25, finance2022 financeNet sales:        finance   Products ! 60,584  ! 63,355  ! 230,901   ! 245,241  finance   Services  21,213    19,604    62,886   58,941  financeTotal net sales (1)  81,797    82,959   293,787    304,182  financeCost of sales:        finance   Products  39,136    41,485    146,696    155,084  finance   Services  6,248    5,589    18,370    16,411  financeTotal cost of sales  45,384   47,074    165,066    171,495  financeGross margin  36,413    35,885   128,721    132,687  finance        financeOperating expenses:        financeResearch and development  7,442    6,797    22,608   19,490  financeSelling, general and administrative  5,973    6,012    18,781    18,654  financeTotal operating expenses  13,415    12,809    41,389    38,144",
        "metadata": {
          "url": "https://www.apple.com/newsroom/pdfs/fy2023-q3/FY23_Q3_Consolidated_Financial_Statements.pdf"
        }
      },
      {
        "collectionId": "finance",
        "documentId": "cc10f2f8-0895-4914-a6d2-1a50b4f679c8",
        "chunkId": "6b4674b5-8419-4f7e-a109-1d96734cb13e",
        "content": "Common stock and additional paid-in capital, !0.00001 par value: 50,400,000 shares financeauthorized; 15,647,868 and 15,943,425 shares issued and outstanding, respectively  70,667    64,849 financeRetained earnings/(Accumulated deficit)  1,408    (3,068) financeAccumulated other comprehensive income/(loss)  (11,801)   (11,109) financeTotal shareholders’ equity  60,274    50,672  financeTotal liabilities and shareholders’ equity ! 335,038  ! 352,755  finance finance finance finance financeApple Inc. financeCONDENSED CONSOLIDATED STATEMENTS OF CASH FLOWS (Unaudited) finance(In millions) finance finance Nine Months Ended finance July 1, finance2023  June 25, finance2022 financeCash, cash equivalents and restricted cash, beginning balances ! 24,977  ! 35,929 finance    financeOperating activities:    financeNet income  74,039    79,082  financeAdjustments to reconcile net income to cash generated by operating activities:    financeDepreciation and amortization  8,866   8,239  financeShare-based compensation expense  8,208    6,760  financeOther  (1,651)   2,695  financeChanges in operating assets and liabilities:",
        "metadata": {
          "url": "https://www.apple.com/newsroom/pdfs/fy2023-q3/FY23_Q3_Consolidated_Financial_Statements.pdf"
        }
      }
    ]
  }
}
Built for Developers
We've got your back

We are developers ourselves and want to provide the best developer experience. Intuitive APIs, stellar documentation, and sample code to get you up and running fast. Support when you need it from people who speak your language.

Your wins are our wins. We want to make you successful, because when you build incredible things, we all grow stronger. The developer community propels the AI ecosystem forward.

So bring your ideas, creativity, and drive. With documentAI as your copilot, let's show the world what we can create together.

Use Cases
See how our customers are using
document
AI
There are many ways to leverage documentAI but here are few case studies to show of the product.