AppsAutessaDB

Semantic search and AI · AutessaDB

Semantic (vector) search

AutessaDB has built-in vector search. Mark a Short Text, Long Text, File or Audio column for semantic search and every insert and update is embedded in the background, with no re-index jobs and no separate vector database. An AI step enriches each value with a summary, key facts and the questions someone would ask of it, so search matches meaning, not just words.

Goes beyond a typical managed database
call_notes · Semantic Searchcolumn: summarySAMPLE
which calls mention pricing complaints?
#7712Customer said the new tier feels too expensive for a team of four0.88
#7690Asked whether annual billing lowers the per-seat cost0.81
#7655Comparing us with a cheaper tool before renewal0.74
Enriched on write: summary · key facts · likely questions. Filter alongside: status = 'open'.

-- the problem

A separate vector store drifts out of sync with the data

Most teams bolt a vector database onto their main database and write a pipeline to keep them in step. Every update is a chance for the index to fall behind, and permissions live in two places. In AutessaDB, the index is a property of the column: the same row, the same security, the same write.

What you get

  • Turn on Semantic Search per column; inserts and updates are indexed automatically.
  • AI enrichment adds summaries, key facts and likely questions to each value before it is embedded.
  • The same table answers "which calls mention pricing complaints?" and status = 'open'.
  • A Semantic Search tab in the builder shows ranked results with similarity scores.
  • Search requests take a query, a column, a limit (default 10, max 100), a scoreThreshold from 0 to 1, a page and includeDetails.
  • Per-column indexing metadata shows what was vectorized and when.
  • Search respects row rules and refuses columns the caller sees only masked or hidden.
  • Bulk writes can index immediately, defer indexing or skip it.

-- how it works

How does semantic (vector) search work in AutessaDB?

  1. Mark the column

    In the column editor, turn on Semantic Search for a text, file or audio column (vectorizeField in the API).

  2. Write rows

    Every write is enriched and indexed in the background.

  3. Search

    Use the Semantic Search tab, query semantic-search in the CLI, the /search API, or give an agent a search tool.

  4. Combine

    Filter structured columns and search meaning in the same workflow, on the same rows.

-- every option

Search options

optionwhat it does
Vectorizable typesSHORT_TEXT, LONG_TEXT, FILE and AUDIO.
query / columnNameThe question and the vectorized column to search.
limitDefault 10, maximum 100.
scoreThreshold0.0 to 1.0 minimum similarity; default 0.
page1-based paging through results.
includeDetailsReturn detailed match information.
Rate limitSemantic searches are rate-limited per company.

-- example

Vector search from the CLIbash
# vector search over a vectorized column
autessa autessadb query semantic-search --tableId ayXduiaLeFKh --json '{...}'

# semantic-search needs a column created with "vectorizeField": true

Full reference in the AutessaDB CLI docs and the API reference.

-- permissions

Who can do what?

accesscan
READRun semantic search on the table.
ADMINTurn semantic search on for a column.

◆ AI assist

Where AI helps

Indexing itself uses AI: each value is enriched with summaries, key facts and the questions someone would likely ask before it is embedded. Search agents query vectorized columns at run time.

AI in AutessaDB works inside the same access rules as the person using it, and schema changes always wait for a person to apply them.

-- relations

How it connects

autessadb feature

File and audio search

Notes and transcripts make media searchable.

autessadb feature

Knowledge bases

The one-click preset for bulk documents.

autessadb feature

Agent tools

Semantic Query and General Search tool types for RAG.

autessa platform

AI agents

Search agents read your tables at run time.

-- faq

Questions about semantic (vector) search

Do I need a separate vector database with AutessaDB?

No. Vector search is a column setting. The embedding index lives with the row, follows its security, and updates on every write.

Which columns can be searched by meaning?

Short Text, Long Text, File and Audio columns marked for semantic search.

Does semantic search respect row-level security?

Yes. Results include only rows the caller can see, and searching a column the caller sees only masked or hidden is refused.

Is it only for knowledge bases?

No. Any vectorized column of any table can be searched.

$ autessa autessadb tables create

Try semantic (vector) search in AutessaDB

Mark a column for semantic search and every write is indexed in the background, enriched by AI, and searchable by meaning next to ordinary filters.

Read the AutessaDB guide or install the CLI.

100% customizable, code and all

Make AutessaDB fully your own.

Use it out of the box, then change anything. Ask Autobot, Autessa’s AI builder, for a new field, screen, report, workflow or agent and it builds it into your copy of the app. Or open the code and edit it yourself. Nothing is locked: the app, its data and its code are yours.

Start free and make it yours