The RAG in 60 seconds sandbox builds a temporary index that deletes itself after 3 days. This page shows how to do the exact same thing, and much more, in your own account, where the index is permanent and private.
You can do every step two ways: by clicking in your index panel (no code), or by calling the API from your app or script. Both are shown below.
The journey, in four steps
Everything after step 1 happens in one place. Open your index from My Indexes and you land in its panel, with the menu on the left:
1Create your index
One index holds your documents and makes them searchable by keyword and by meaning (vectors). Create it once.
Open My Indexes and click:
Name it (for example my_research), pick the region closest to you, confirm. Live in seconds, with vector search already on.
POST https://opensolr.com/solr_manager/api/create_index
email=YOUR_EMAIL & api_key=YOUR_API_KEY & solr_index_name=my_research2Add your data
Three ways to fill your index. Pick the one that matches what you have.
| You have | Use | Where |
|---|---|---|
| A website | WebCrawler (no code) | Left menu › WebCrawler |
| PDFs, DOCX, document URLs | Data Ingestion (auto text + vectors) | API, watched on your Data Ingestion page |
| A few JSON records | Solr Admin (paste by hand) | Left menu › Solr Admin |
Open your index, click WebCrawler in the left menu, paste the site address and start it. Opensolr fetches the pages, extracts the text and builds the vectors for you.
Send a document URL (PDF, DOCX, web page) or inline JSON. Opensolr downloads it, pulls the text, even runs OCR on scans, and embeds every document automatically.
POST https://api.opensolr.com/solr_manager/api/ingest
{"email":"YOUR_EMAIL","api_key":"YOUR_API_KEY","core_name":"my_research",
"documents":[{"uri":"https://example.com/report.pdf","title":"Report","rtf":true}]}Set rtf:true to extract the file text. Watch progress on your Data Ingestion page, or by API:
GET https://api.opensolr.com/solr_manager/api/ingest_status?email=YOUR_EMAIL&api_key=YOUR_API_KEY&job_id=JOB_IDFor a handful of records, open Solr Admin in the left menu and paste JSON documents directly. Good for testing, not for bulk.
3Open your search page
Every index has a ready hosted search page. Two ways: open the page, or call the search API from your app.
Your index gets its own page at the address below. Open it and type a question in any language, for hybrid results (keywords plus meaning).
You can password‑protect the page from the panel, and drop it on your own website with the one‑line embed.js widget.
One call for hybrid search:
POST https://api.opensolr.com/solr_manager/api/embed_and_searchTwo calls for a full RAG answer with citations, exactly like the sandbox: embed_and_search finds the passages, then ai_summary writes the answer.
4Turn on the chat bot
Give your index a chat assistant that answers from your documents, with citations. It lives on your site as a small widget.
Open your index and click Chat Bot in the left menu:
Your account or the API: which one?
| Your account | API | |
|---|---|---|
| Best for | Doing it by hand, once | Your app, script or site, automated |
| Create index | Create New Index button | create_index |
| Add data | WebCrawler, or Solr Admin | ingest |
| Search | Hosted search page | embed_and_search + ai_summary |
| Chat bot | Chat Bot menu, Enable | assistant_chat |
The result is the same index either way. Click through it first to understand it, then switch to the API inside your own app.
More detail: testing your AI search · vector embeddings · crawler API · crawler vs ingestion