nocobase

Provides a drag-and-drop interface for designing database tables, fields, and relationships without writing SQL, with bidirectional synchronization between schema definition and UI rendering. Uses a declarative schema model that maps to underlying relational databases (PostgreSQL, MySQL, SQLite) and automatically generates CRUD forms, tables, and views from field definitions. Changes to schema immediately reflect in the UI layer through event-driven updates.

Accepts natural language descriptions of business processes or data requirements and generates database schemas, field configurations, and workflow automation rules through LLM integration. The system parses user intent, maps entities and relationships, and scaffolds initial table structures, field types, and automation rules that users can then refine in the WYSIWYG builder. Supports multiple LLM providers (OpenAI, Anthropic, local models) via a pluggable provider interface.

Allows users to define webhooks that are triggered by database events (record creation, update, deletion) and send HTTP POST requests to external systems with record data as payload. Supports webhook filtering (only trigger on certain field changes), retry logic with exponential backoff, and webhook signature verification for security. Webhook delivery status and logs are stored for debugging.

Provides pre-built application templates (CRM, project management, inventory, HR) that users can instantiate with a single click. Templates include schema definitions, views, workflows, and sample data. Users can customize templates after instantiation or create their own templates from existing applications. Templates are stored as JSON and can be shared across NocoBase instances.

Supports importing data from CSV, Excel, JSON, and other formats with automatic schema detection and field mapping. Users can preview data before import, define field mappings, and handle data type conversions. Supports bulk export to CSV, Excel, and JSON formats with field selection and filtering. Import/export operations are logged and can be scheduled.

Allows users to define computed fields using a formula language that supports arithmetic, string operations, date functions, and conditional logic. Formulas are evaluated on-demand or cached depending on configuration. Supports referencing other fields in the same record and related records through relationships. Formula syntax is similar to Excel with functions like SUM, CONCAT, IF, etc.

Provides a visual workflow builder for defining automation rules triggered by database events (record creation, field updates, deletion) or external webhooks. Rules support conditional branching (if-then-else), loops, and actions including field updates, API calls, email notifications, and plugin-based custom actions. Workflows are stored as JSON configuration and executed by a server-side engine that maintains execution state and logs.

Automatically generates multiple presentation views of the same underlying data without duplicating data or creating separate tables. Supports grid (spreadsheet-like), form (record detail), kanban (column-based grouping), calendar (date-based), and gallery (card-based) views. Each view is configurable with filters, sorting, grouping, and field visibility settings. Views are stored as metadata and rendered on-demand from the same data source.

Implements a hierarchical permission system where access is controlled at multiple levels: table (read/write/delete), field (visibility and editability), and record (row-level security based on user attributes or record values). Permissions are defined through a UI and stored in the database, evaluated at query time to filter data and at render time to show/hide UI elements. Supports role inheritance and dynamic permission rules based on user properties.

Automatically generates REST and GraphQL APIs from the database schema and field definitions without requiring manual endpoint definition. Each table gets standard CRUD endpoints (/api/tables/{id}/records with GET, POST, PUT, DELETE), with query parameters for filtering, sorting, pagination, and field selection. GraphQL schema is generated from table relationships and field types. APIs respect permission rules defined in the access control system.

Provides a plugin architecture allowing developers to extend NocoBase with custom field types, view components, workflow actions, and API endpoints. Plugins are TypeScript/JavaScript modules that register with the plugin registry and hook into the application lifecycle. The system supports both server-side plugins (for custom actions and API endpoints) and client-side plugins (for custom views and field components). Plugins can access the database, permission system, and workflow engine.

Provides a file attachment field type that stores files in configurable backends (local filesystem, AWS S3, Aliyun OSS, MinIO) with automatic metadata extraction (size, MIME type, dimensions for images). Files are referenced in database records through attachment objects containing URLs, metadata, and access control information. Supports batch upload, drag-and-drop, and direct URL linking. File access respects record-level permissions.

Implements real-time data synchronization across multiple users editing the same records simultaneously using WebSocket connections and operational transformation (OT) or conflict-free replicated data types (CRDTs). Changes made by one user are broadcast to others in near real-time, with automatic conflict resolution when multiple users edit the same field. Supports undo/redo with proper conflict handling.

Automatically logs all changes to records including who made the change, when, what field was modified, and the old/new values. Maintains a complete audit trail that can be queried and filtered. Supports rollback to previous versions and change notifications. Audit logs are stored in a separate table and respect the same permission rules as the original data.