Skip to content
  • App
  • Make
  • n8n

How to answer questions from your documents

To answer questions from your own documents, add the documents to a 1001fx knowledge base with rag/adddocument and ask it with rag/ask – in Make, in n8n or in the 1001fx app. 1001fx stores the text, finds the passages that fit a question and lets AI answer from them, with the sources – no vector database and no AI account needed.

Get your free API keyTry it in the appadddocument reference

The fields

The same in the 1001fx app, Make and n8n. adddocument costs 1 credit per 1,000 tokens (minimum 1).

knowledgeBase
– The name of your knowledge base, for example Support FAQ – it is created on first use. In Make you can also pick it from a drop-down.
file
– The file from the previous step (PDF, Word, PowerPoint, Excel, HTML, text, Markdown or CSV).
documentId
– The Google Drive file ID. The same ID replaces the document when the file changes, so you never create duplicates.
title
– The file name, shown in the sources of an answer.
sourceUrl
– The web view link of the file, so the sources link back to Google Drive.
useOcr
– Yes for scanned PDFs and images without a text layer (6 credits per page extra).

All parameters and examples: adddocument reference.

Example

What you send to adddocument and what comes back – the same in every tool.

RAG – Add Document to Knowledge Base

Needs an API key of an account (not a test key).

Input

Expects a json body
{
  "knowledgeBase": "Support FAQ",
  "documentId": "faq-billing",
  "title": "FAQ Billing",
  "text": "How do I cancel my subscription? Go to Settings → Subscription and click Cancel.",
  "metadata": {
    "department": "billing"
  }
}

Output · HTTP 200

Response
{
  "statusCode": 200
}
Show curl command
curl -X POST 'https://api.1001fx.com/rag/adddocument' \
  -H 'X-API-KEY: YOURSECRETAPIKEY'

In the 1001fx app

Without a scenario or workflow: run the function in your browser for a one-off task, or to try the fields before you build the automation.

adddocument in the 1001fx app: the form with example values
RAG – Add Document to Knowledge Base in the 1001fx app
  1. 1

    Open the function

    Log in to the 1001fx app and open Functions → RAG – Add Document to Knowledge Base.

  2. 2

    Fill the fields

    Fill the fields as described above. Files can be uploaded from your computer.

  3. 3

    Run it

    Click Run. The answer appears next to the form; files come as a download. Each run is charged like an API call.

  4. 4

    Reuse the call

    The same page shows the call as cURL command, ready for scripts and other tools.

In Make

  1. 1

    Start the scenario

    Start with the module that delivers new or changed files, for example Google Drive "Watch files in a folder" followed by "Download a file". A second scenario answers the questions, for example Slack "Watch public channel messages" → 1001fx "RAG – Ask Knowledge Base" → Slack "Create a message".

  2. 2

    Add the 1001fx module

    Add a module, search for 1001fx and pick “RAG – Add Document to Knowledge Base”. Select your 1001fx connection or create one with your API key.

  3. 3

    Fill the fields

    Fill the fields as described above, mapped from the previous steps.

  4. 4

    Use the result

    The answer contains the stored document with its passage count and the tokens read. In the question scenario, map the answer field of RAG – Ask Knowledge Base into your reply and list the titles and source URLs of the sources below it.

Connecting 1001fx: 1001fx for Make.

In n8n

  1. 1

    Start the workflow

    Start with a Google Drive Trigger node that watches a folder for new or changed files, followed by a Google Drive node that downloads the file. A second workflow answers the questions, for example Slack Trigger → HTTP Request to rag/ask → Slack node that posts the answer.

  2. 2

    Add the HTTP Request node

    Add the HTTP Request node and send a POST request to https://api.1001fx.com/rag/adddocument with the header X-API-KEY and your API key.

  3. 3

    Fill the fields

    Fill the fields as described above, mapped from the previous steps.

  4. 4

    Use the result

    The answer contains the stored document with its passage count and the tokens read. In the question workflow, use the answer field of the rag/ask response in the Slack node and list the titles and source URLs of the sources below it.

How to set up the HTTP Request node: 1001fx for n8n.

Tips

  • Unchanged files are not read again and cost 0 credits, so the scenario or workflow can safely run on every change.
  • Add metadata such as department or product and filter on it in rag/ask, so one knowledge base serves several teams.
  • Remove documents with rag/deletedocument when the file is deleted in Drive, so answers never quote old content.
  • For large files, send the call with the header x-fx-webhook so it runs in the background.

Frequently asked questions

Do I need a vector database or an OpenAI account for RAG in Make or n8n?
No. 1001fx stores the documents, creates the embeddings, searches and writes the answer. Make and n8n only need your 1001fx API key.
What does it cost?
Adding costs 1 credit per 1,000 tokens read, a question 4 credits per started 1,000 tokens, storage 1 credit per MB and month with 10 MB free.
What happens when the answer is not in the documents?
The AI says that the answer is not in the documents instead of guessing.
Ask 1001fx