Search by meaning & answer from your documents

Let your app understand what people mean, not just the exact words they type, and answer questions from your own documents and content.

This feature requires the Pro plan or higher.

What It Does

Most search only matches the exact words people type. This feature lets your app search by meaning, so a question like "how do I get my money back?" still finds your "refund policy" page even though none of the words match. It is what people mean by semantic search.

You give your app the content it should know - support articles, FAQs, product docs, uploaded files - and it can search that content by meaning and answer questions from it. Behind the scenes, GenMB turns your text into a form the computer can compare by meaning (this is called embedding), and stores it so searches are fast. There is no AI model to choose and no API key to manage; it is all handled for you.

Search your content by meaning, not just exact keyword matches.
Add raw text or upload files - PDF, TXT, MD, and CSV are supported.
Long documents are split into overlapping passages automatically, so each piece keeps enough context.
Tag content (for example by category or language) and narrow searches to those tags.
No AI model to pick and no API key to manage - embedding is fully managed.

How to Add It

The fastest way to add it is to ask in chat, for example "let visitors search our help articles by meaning" or "add a Q&A box that answers from the PDFs people upload." GenMB wires the content and search into your app for you, and you can refine the wording and placement from there.

Give your app the content

Add the text, articles, or files your app should know about. Long documents are split into searchable passages automatically.

Let people search by meaning

Searches return the most relevant passages, ranked by how closely they match the meaning of the question, along with the document each passage came from.

Keep your content current

Review what is stored and remove documents you no longer need at any time.

Answer Questions From Your Content

Pair this with the AI Chatbot to build a question-and-answer assistant grounded in your own content. The app finds the passages most relevant to a question, then the chatbot answers using only those passages instead of guessing. This approach is often called RAG (retrieval-augmented generation): retrieve the right content first, then let the AI generate an answer from it.

Show where answers come from

Show the source document alongside each answer so people can click through and check it. That is what turns a generic chatbot into a trusted assistant people rely on.

When to Use It

Reach for this whenever search by meaning matters more than an exact keyword match.

Knowledge bases - internal docs, FAQs, support runbooks, product handbooks.
Customer support assistants that need to answer from your own content.
Document Q&A apps - let people ask questions across uploaded PDFs.
Recommendation features - find related articles, products, or items by how similar their descriptions are.

For structured business data - users, orders, projects - use the relational Database & Schema feature instead. The two complement each other in the same app.

Quotas & Limits

Limits are counted per app. The hourly limits below reset each hour.

Limit	Pro	Business
Max documents	10,000	100,000
Max storage	1 GB	10 GB
Searches per hour	1,000	10,000
New documents added per hour	100	1,000

For Developers

The SDK is exposed on window.genmb.vectordb when the capability is enabled in your app. It is backed by pgvector (a vector search extension) on the shared Cloud SQL instance, with no external vector store to wire up. Embedding uses Gemini text-embedding-004 (768 dimensions).

Add content

Call ingest(text, metadata) with raw text, or ingestFile(file, metadata) with a File from an upload input. Returns { documentId, chunks }.

Search by meaning

Call search(query, { limit, filter }). Each result includes content, a similarity score, the original metadata, and the parent documentId. A filter on the metadata is applied before similarity ranking.

List or delete documents

Use list() for an admin view or delete(documentId) to remove a document and its chunks.

To answer questions from your content, combine search with the AI Chatbot SDK: retrieve the most relevant passages, then pass them as context.

const { results } = await window.genmb.vectordb.search(userQuestion, { limit: 3 }) const context = results.map(r => r.content).join('\n\n') const answer = await window.genmb.chatbot.send( `Answer based on this context:\n${context}\n\nQuestion: ${userQuestion}`, { systemPrompt: 'Answer based only on the provided context. If the answer is not in the context, say so.' } )

FAQs

How is this different from the Database & Schema feature?▾

The Database & Schema feature gives you tidy tables for structured business data - users, products, orders. This feature stores your documents and text so the app can search them by meaning. They work together: keep your business data in the database, keep your knowledge content here.

Do I have to set up an AI model or API key?▾

No. Turning your content into searchable meaning (the "embedding" step) is fully managed for you. Under the hood GenMB uses Google Gemini's text-embedding-004 model (768 dimensions) for both adding content and searching it.

Do I need to split long documents into pieces myself?▾

No. Long documents are split into overlapping passages automatically when you add them, so each searchable piece keeps enough context to make sense on its own.

Can I narrow searches to a category or language?▾

Yes. You can tag content when you add it (for example category "support", language "en"), then scope a search to only those tags. The app narrows to the matching content first, then ranks it by meaning.

What happens when I reach the document limit?▾

Adding new content stops working until you remove some documents or upgrade your plan. You can review what is stored and delete individual documents at any time. Limits are counted per app, not per person.

Ready to build?

Create your first app for free, no credit card required.

Database & Schema Data Connectors