-- every feature
Every AutessaDB feature
AutessaDB is the database for AI apps and agents, ready out of the box and shaped to how your company works. Here is everything it does, grouped the way you would build with it: model the data, query it, search it by meaning, secure it row by row and column by column, automate it with triggers, analyze billions of rows in the data lake, and run files and apps on top of it.
-- 01 · data_modeling
Tables and data modeling
AutessaDB tables are typed, related and computed: text, number, select and media columns, foreign keys, AutessaScript computed columns, write-once ledger tables and one-click knowledge bases.
Tables and column types
AutessaDB tables have typed columns from Short Text to Audio, with required flags, value limits, select lists and system columns on every row.
Read more →Foreign keys and joins
Integer columns can be foreign keys to rows in another table, and join queries follow them with filters, while each joined table keeps its own security.
Read more →Computed columns
A computed column runs a short script over the row after each write and stores the answer, recomputing only when the columns it depends on change.
Read more →Append-only tables
Append-only tables are write-once: rows can be added and read but never changed or removed, and the setting can be turned on but never off.
Read more →Knowledge bases
Drag files or whole folders into a knowledge base and they are indexed for semantic search, with a Create Agent shortcut for a ready-made search agent.
Read more →-- 02 · queries_and_apis
Queries, APIs and SDKs
One query model everywhere: the same filters, sorts, aggregates and bulk writes work in the builder, the REST API, the CLI, scripts and agent tools.
Filters, sorting and pagination
Queries take groups of AND/OR conditions with comparison, text, list and null operators, multi-column sorts, and either page numbers or cursors for large tables.
Read more →Aggregations
Group rows by a few columns and compute COUNT, SUM, AVG, MIN and MAX on the server, filtering the groups with HAVING.
Read more →Bulk and batch writes
Bulk calls write many rows by id and batch calls change every row matching a filter, with control over atomicity, triggers, indexing and computed columns.
Read more →REST API, CLI and SDK
The REST API, the `autessa autessadb` CLI namespace and the `autessa.autessadb` SDK reach the same tables with the same security as the builder.
Read more →-- 03 · vector_search
Semantic search and AI
Vector search is a column setting, not a second database: flip it on and AutessaDB indexes, enriches and serves every value to people and agents.
Semantic (vector) search
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 more →File and audio search
File columns get an auto-written notes companion and audio columns get a transcript, so documents and calls are searchable like text.
Read more →Database tools for AI agents
Ready-made tools give agents insert, update, delete, query, aggregate, semantic query, general search, Drive and data lake access to your tables.
Read more →DB Copilot
Describe the data you need and the DB Copilot proposes tables and schema changes as cards you apply or reject, and answers questions about your data.
Read more →-- 04 · row_level_security
Row-level security and access
Every read and write, from a person, agent, script, app or API key, passes through the same row rules, column masks and access log.
Row-level security
Row rules decide which rows each caller sees, with placeholders for the current user, teams, roles and app visitors, and edit checks for changes.
Read more →Column masking and hidden columns
Per role, a column is visible, masked (fully redacted, last four, or a consistent code) or hidden, and masked columns are stored encrypted.
Read more →Security roles
A role bundles table policies (row rule, edit check, column rules) and is assigned to people, teams, company roles, everyone, custom apps or API keys.
Read more →Sharing and access levels
Tables start private and are shared at READ, WRITE, PUBLISH or ADMIN, with a default security role applied to everyone you share with.
Read more →Data access log
Every read, change and reveal of protected data is logged by table and row, with a chain integrity check and filters by date and action.
Read more →-- 05 · triggers
Triggers and pipelines
Row changes become automation: triggers run tools, agents and writes, chain across tables, and leave a full execution history.
Database triggers
A trigger fires on insert, update or delete, checks column conditions, and runs a tool, an agent, a multi-agent or a table update with the row values.
Read more →Pipelines and impact analysis
Triggers that write rows into other tables chain into pipelines, and an impact analyzer maps the cascade and warns about cycles before anything runs.
Read more →Trigger execution history
Every firing keeps a snapshot of the trigger, the row before and after, and the action response, so debugging starts from facts.
Read more →-- 06 · data_lake
Data lake and analytics
A columnar data lake for events, logs, history and exports: bulk loads, partitions, analytical SQL, a semantic layer, plain-language Ask, data agents and archive tiers, all under the same security as your tables.
Columnar data lake tables
Lakehouse-style columnar tables for events, logs, history and exports, built for aggregations, group-bys, window functions and joins over billions of rows.
Read more →Bulk loading and partitioning
Load CSV and Parquet files or whole AutessaDB tables, append or replace, see every bad row reported, and partition by date so queries read only what they need.
Read more →SQL Workbench
A SQL editor that knows your tables and measures, estimates scan size as you type, keeps history, and downloads results as CSV or Parquet.
Read more →Semantic layer
Descriptions, grain, synonyms, measures, joins and verified queries teach people and agents how to query a table, and Suggest with AI drafts them for review.
Read more →Ask in plain language
Type a question such as "monthly revenue by region this year" and Ask proposes SQL from your descriptions, measures and verified queries. Nothing runs until you click Run.
Read more →Data agents
One click creates describe, query and result tools for the tables you pick, so an agent can answer data questions with SQL, preferring your verified queries.
Read more →Security on every query
Sharing, row rules, hidden columns and masked columns apply to every SQL query, and masked values still group, count and join without ever being shown.
Read more →Retention and archive tiers
Retention deletes rows past a set age; the archive tier moves older days to compressed Parquet and restores them as a read-only table when you need them.
Read more →-- 07 · files_and_apps
Files, apps and operations
AutessaDB is also the backbone for files and apps: Drives, app sign-in tables and per-visitor rules, with soft delete as the safety net.
Drive on AutessaDB
A Drive is an AutessaDB table of files and folders, so files get the same access rules, agent tools, CLI and SDK as every other table.
Read more →Backend for custom apps
Custom apps read and write AutessaDB tables, keep their sign-in users in AutessaDB, and scope each visitor to their own rows with role placeholders.
Read more →Soft delete and restore
Table deletion needs a confirmation token, lands in a trash with 14 days to restore, and ownership can move to a teammate.
Read more →-- full capability list
- Typed tables with Short Text, Long Text, File, Image, Video, Audio, Integer, Decimal, Boolean, Single Select and Multi Select columns
- System columns id, created_at and last updated on every row
- Required flags, length and value limits, and select option lists
- Foreign keys between tables and join queries
- Computed columns written in AutessaScript, with dependency ordering
- Append-only (write-once) tables for ledgers and audit data
- Knowledge bases: drag-and-drop file ingestion ready for semantic search
- Semantic (vector) search on text, file and audio columns, indexed automatically on every write
- AI enrichment of indexed values with summaries, key facts and likely questions
- Auto-generated notes for file columns and transcripts for audio columns
- Structured queries with grouped AND/OR filters and a full set of comparison, text, list and null operators
- Sorting, offset paging and cursor (keyset) pagination
- Group-by aggregates: COUNT, SUM, AVG, MIN, MAX with HAVING
- Bulk insert, update and delete by id, and batch update and delete by filter
- Atomic or best-effort bulk writes with trigger, indexing and compute modes
- Indexed columns for fast equality filters
- REST API, Autessa CLI and AutessaScript SDK over the same tables
- AutessaDB tool types so agents insert, update, query, aggregate and search tables
- DB Copilot: design schemas and query data in plain language, with reviewable change cards
- Row-level security roles with row rules and user, team, role and app placeholders
- Edit checks that stop people moving rows outside what they can see
- Column masking: visible, masked (redact fully, last 4, hash) or hidden per role
- Masked columns stored encrypted, with every reveal recorded
- Roles assigned to people, teams, company roles, the whole company, custom apps and API keys
- Sharing at READ, WRITE, PUBLISH and ADMIN levels
- Row-level data access log with a chain integrity check
- Triggers on INSERT, UPDATE and DELETE that run tools, agents, multi-agents or table updates
- Trigger conditions: changed, became null, became non-null, comparisons, text and multi-select operators
- Trigger chains as multi-stage pipelines, with a design-time impact and cycle analyzer
- Trigger execution history with before and after row snapshots and responses
- Columnar data lake for millions to billions of rows: events, logs, history and exports
- Lakehouse-style columnar storage you query with analytical SQL, on the same platform as your relational tables and Drive
- Bulk loads from CSV and Parquet uploads, server-fetched files and existing AutessaDB tables, append or replace, with bad rows reported
- Partitioning by year, month, day, hour, identity, bucket or truncate, with a scan estimate before every query
- Read-only ANSI SQL with joins, CTEs, subqueries, unions and window functions
- SQL Workbench with schema-aware completion, estimates, history and CSV or Parquet downloads
- Queries run as asynchronous jobs you can wait on, check later, cancel or download
- Semantic layer: descriptions, grain, synonyms, measures, joins and verified queries, drafted with AI
- Ask in plain language: a question becomes proposed SQL you review and run
- Data agents that answer questions with SQL through generated describe, query and result tools
- Dashboards in custom apps over the same read-only queries
- Masked columns in the data lake that still group, count and join
- Row- and column-level security and sharing applied to every SQL query
- Data lake retention policies and archive tiers that move older days to compressed Parquet, with restore
- Drive: file drives built on AutessaDB with folders, versions and search
- Backend for custom apps, including sign-in user tables and per-visitor row rules
- Soft delete with a 14-day restore window and two-step table deletion
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