Add and Analyse Your Own Text in RAG in 60 Seconds

On RAG in 60 seconds you can add your own text, a paragraph, an article or a whole page, and then search it, ask it questions and see what Opensolr worked out from it. This page shows where to put the text and what you get back.

1. Create your sandbox index

Click Create my sandbox index. You get your own Opensolr Index with 100 sample documents already in it. It lives 3 days. Keep the shareable link: it is the only way back to it.

2. Add your text

Under Push some data, pick one of the tabs:

Paste JSON: for text you have in hand. Replace the example with your own text and click Ingest JSON. Up to 50 texts at once.
From a URL: for text that is already on a web page. Give the page, your homepage or your sitemap.xml, and click Start crawling my site. Up to 100 pages are read.
curl: the same thing from your own terminal, with your sandbox credentials already filled in.

For one text, paste this into Paste JSON and put your text in text:

[
  {
    "uri": "https://example.com/my-text",
    "title": "My text",
    "description": "A short summary of my text.",
    "text": "Put the text you want to analyse here. One sentence or many paragraphs."
  }
]

uri only has to be a web address that is unique for each text; it does not have to exist. Each text needs all four fields. For a PDF, see the note under the Paste JSON tab.

3. Wait for the queue

Watch the ingestion queue shows your job. When it says completed, your text is in the index, usually within seconds.

4. See what Opensolr made of it

Ask your data: ask a question about your text; the answer is written from it, with its source.
Prove the hybrid: search your text by its exact words, then by describing it in other words, and compare keyword, meaning and hybrid results.
Browse the raw Solr index: open your document and see what was added to it: the language, the sentiment and the meaning vector of your text.
The search page at the bottom: search it like a visitor of your site would, with AI Hints and the AI Reader.

Only want the analysis, without an index?

To get the language, sentiment and meaning vector of a text straight back, with nothing stored, use the enrich_text call of the AI API with your own account: Language, sentiment and vectors for your texts.

To put your own text in an index of your own account, for good: Your first push.