Feature overview|V0.25.0|OceanBase DataPilot|docs|OceanBase Documentations

This topic lists the features currently provided by DataPilot and their corresponding documentation links, organized by chapter, to help you quickly understand its capabilities. You can use this table to confirm whether a feature is available during evaluation and selection, or to directly jump to the operation documentation for daily use.

Whether a feature is available also depends on the current role, workspace membership, detailed permissions, and object release status. The absence of an entry in the interface does not mean the feature does not exist. For troubleshooting, refer to the FAQ.

Feature matrix

Deployment and initialization

Feature Sub-module	Feature	Description	Documentation
Platform Deployment	Tool Deployment, High Availability Deployment, and OAT Deployment	Select a deployment method based on your organizational environment, complete platform initialization, and create an environment where you can log in and create workspaces.	Tool deployment, high-availability deployment, and OAT deployment
License	Import, activate, view validity period and capacity	Declares the authorization boundary; when capacity reaches the upper limit, objects must be cleaned up first or the license must be updated.	License Activation and Capacity Management
Running Basics	Sandbox, Endpoint, and External Service Configuration	Maintain platform runtime dependencies. After saving, verify availability through actual business operations.	Endpoint Registration and Sandbox Configuration
Model Management	Vendor credentials, model instances, and connectivity probing	Maintain available models and select by model type, capability tag, and maximum context	Connect to Model Service Provider
Model Management	Business module binding, strategy considerations, and rate limiting	Bind models and strategies to text indexing, semantic processing, and general dialogue respectively.	Model registration and module binding

Data management

Feature Sub-module	Feature	Description	Documentation
Data Source Management	Create, edit, test connection, save	A successful connection test only proves connectivity at that moment; a successful save is what persists the configuration.	Create Data Source
Data Source Management	Target data source	To inherit the results of the commit and write objects back, write permission is required.	Configure Target Data Source
Semantic layer	Metadata browsing and data table preview	The data table preview returns at most 10 rows. Permission previews apply policies based on the selected preview identity, while external hosting mode follows the hosting configuration. It is not a full table export.	Metadata browsing and retrieval
Semantic layer	Select Range, Initiate Synchronization, Resume from Breakpoint, and Rerun Entire Process	Synchronization is a long-running background task. You can view its final status under the Semantic category in the Task Center.	semantic processing
Semantic layer	Business Objects, Attributes, Relationships, Metrics, and Filter Conditions	The metric caliber is determined by both the calculation logic and filter conditions and requires manual verification.	Business Definition and Metric Maintenance
Semantic layer	Dimension Value Synchronization and Aliases	Make sure the agent understands the business terminology. Before clearing an alias, confirm that there are no dialogues or metric dependencies.	Synonym Management
Semantic layer	Ontology Topology, Confidence, Evidence, and Metric Evaluation	Topology is used to verify objects and relationships; confidence represents the certainty of the inference, not accuracy.	Ontology Topology
Data Permission	Member matching, table/row/column limits, local policies, and external hosting	Policies apply uniformly to both agents and data applications. The system supports two modes: local policy management and external data permission management.	Configure Data Permissions
Data Quality	Table lineage and column lineage	The result source is explained. The absence of upstream nodes may be due to permission-related invisibility.	Data Lineage and Data Overview

Intelligent analysis agent

Feature Sub-module	Feature	Description	Documentation
Lifecycle	Create, edit, clone, and delete	Cloning generates an independent replica that does not share subsequent modifications with the source agent.	Create and Configure an Agent
Configuration	Basic Information, Recommended Questions, Data Scope	The data range supports up to 256 items by default, and the effective upper limit for any deployment does not exceed 1,000.	Create and Configure an Agent
Configuration	Clarification, Sub-agent, Maximum Execution Steps and Retry, Rules	Increasing the number of execution steps will increase time consumption and resource usage, and cannot replace proper resource configuration.	Behavior parameters and custom rules
tools	MCP Service Configuration, Sensitive Information Protection, and Restricted Network Target Authorization	Configuring MCP does not mean the current agent or user can call all tools.	Toolset and MCP Access
Asset	Knowledge and Skill Binding	Provide business guidelines and supplementary materials for the agent	Knowledge Base Integration and User Memory
Asset	User Memory	Dialogue recall, start/stop, view, delete, and clear, distinct from agent knowledge and skills	Knowledge Base Integration and User Memory
Asset	QA Evaluation and SOP Learning Loop	Accuracy Report, SOP Learning, and Manual Review for Major Compaction Skills	Knowledge Base Integration and User Memory
Release	Release Verification, Release Status, and Visibility Scope	Publication does not automatically grant data, tool, or subscription permissions.	Version Management and Visibility
market	Discover, Subscribe, Apply, Unsubscribe	Visibility, subscription, and resource access are three independent criteria for judgment.	Agent Market and Subscription
market	Subscriptions formed by subscription approval or external authorization	External authorization directly results in a valid subscription and does not generate an application record within the platform.	Subscription Approval

Dialogue analysis and applications

Feature Sub-module	Feature	Description	Documentation
Dialogue Analysis	Create Session, Stream Answer, Stop	The streaming status indicates that the event stream is still active, not the completion percentage.	Start conversation analysis
Dialogue Analysis	Result Verification, AgentReplay, and Abnormal Calls	Replay is a read-only explanatory view that does not rerun or modify the session.	View and verify analysis results
Dialogue Analysis	Session List, Secondary Tab, and Fork	Branch from the session history list, copy the existing history and session files of the selected session to create an independent session.	Multi-turn Dialogue and Session Management
Dialogue Analysis	File Upload and File Q&A	Successful upload does not guarantee that the answer will use this file. You must explicitly indicate so in your question.	File Upload Analysis
Dialogue Analysis	Feedback, Feedback inbox	Marking as read indicates that the topic has been viewed, but does not mean the configuration has been fixed or re-released.	Feedback Loop
Dashboard	Create a dashboard, configure queries and charts	After generated via dialogue, it can be further structurally edited in the workspace.	Create Dashboard
Dashboard	Filter controls	Control the value range for dashboard queries.	Filter controls
Dashboard	prediction	Provide forecast results for qualified series charts	Intelligent Prediction
Multi-dimensional table	Base tables and derived tables, filtering, sorting, grouping, and file import	File format and size limits are subject to the prompts of the current file selector.	Multidimensional Table Basics
Multi-dimensional table	AI columns, dependency graphs, caching strategies, and runtime scopes	The running scope directly affects costs, time consumption, and result consistency.	AI Column
Multi-dimensional table	Result Persistence, Dependency Chain, and Private Columns	The frozen replica will not automatically follow subsequent configuration changes.	Materialized Data Views and Analyst Personal Columns
Data Dashboard	Generate Dashboards, Bind Data and Materials	Published and Shared with Agent	Create Data Dashboard
Data Dashboard	Canvas Editing and Presentation	For duty display and reporting presentations	Canvas, Materials, and Full-Screen Demo
Notebook	SQL, Python, Markdown, charts, input controls	The execution result can be saved as a snapshot for reuse in downstream applications.	Notebook
report	Generative Reporting and Orchestration Reporting	Unified reporting entry, but different maintenance methods for the two types of reports	Report Overview
report	Generation, Refresh, Versions, and Snapshots	A snapshot represents the state at a specific point in time. Refreshing generates a new version of the facts.	Refresh and Version Management
report	Share and Public Link	The sharing scope should be controlled based on the sensitivity level.	View and share reports
report	Analyst Individual Edition	You can generate a personal version based on the published report without rewriting the publisher's report.	Report Overview
Generative data applications	Requirement generation, dialogue iteration, and preview	Generate interactive applications from requirement descriptions and publish them with the agent.	No special topic available
Export	Table Export, Extraction Result Export, Report Download	The number of rows and the volume limit for each channel are subject to the page prompts.	Data Export

Automation and knowledge

Feature Sub-module	Feature	Description	Documentation
Workflow	Node Orchestration, Connection, and Parameter Configuration	Node types and configurable parameters are based on the current canvas.	Node Reference
Workflow	Run the test and view the logs	Passing the test does not guarantee that the release version will be successful.	Testing and Running
Workflow	Version Take Effect, Scheduling, Enable and Disable	After saving and testing the workflow, you can publish the Agent to make the new version effective. Published workflows can be triggered by authorized users, and their final status can be viewed in the Task Center.	Version and Enablement Management
Knowledge Management	Document Import, External Knowledge Sources	The format, quantity, and size limits are subject to the prompts of the current selector.	Knowledge Documentation and External Knowledge Sources
Knowledge Management	Index, Tag	The index creation is a long-running background task. You can view the failure reasons in the Task Center.	Index and Tag Management
Knowledge Management	Retrieval and Recall	Search results can be traced back to specific document fragments	Knowledge Retrieval
Knowledge Management	Knowledge Assessment	Verify the agent response quality using Q&A test cases and expected answers.	Knowledge Assessment
Intelligent Data Processing	Extract template configuration, test, batch process, and store results	Supports CSV, XML, images, and PDF formats; results can be stored in a database or written to a knowledge base.	Extraction Factory
Intelligent Data Processing	Categorical datasets, model training, batch inference, and sample augmentation	Sample data after review can be re-added to the training set	Document classifier
Intelligent Data Processing	Map analysis, map provider, rendering status	Built on real data and released with the Agent	Spatio-temporal analysis

System management and developer integration

Feature Sub-module	Feature	Description	Documentation
Identity Management	User creation, activation, batch import, activation code	Batch actions are asynchronous management operations and do not generate independent task details.	User and Authentication Management
Identity Management	Login integration, automatic account creation, session validity period	The session validity period applies to subsequently issued credentials and does not affect the current credential.	User and Authentication Management
Permission Management	System Roles, Organization and User Groups, Space Roles	Being able to log in does not mean you can access the workspace or Agent.	Role and Permission Management
System Configuration	Platform parameters and external services	After successful saving, verify through the business path.	System Configuration
O&M Observability	Export system logs and error traces	The exported materials may contain sensitive context and must be stored in a controlled location.	O&M Log
O&M Observability	Task Center: Semantics, Knowledge Index, AI Columns, Workflow, Batch Testing	Observation entry does not replace the source module for modifying business configurations.	Task Center
O&M Observability	Monitor health and sessions, audit operations	Successful audit does not guarantee the success of subsequent asynchronous tasks.	Monitoring Center and Audit Logs
Developer Integration	OpenAPI、API Key	Manage your API key as a secret and do not paste it into documents, screenshots, or conversations.	Open APIs
Developer Integration	Embed and Share	Embedding does not expand the permission scope of the embedded object.	Embed and share

Quick reference by role

Data engineer

Primary Tasks	Entry	Documentation
Access data and confirm visibility scope	Workspace Settings > Data Sources	Create Data Source
Build the semantic layer and validate the metrics	Agent Workspace > Semantic Processing	semantic processing
Create, verify, and publish the agent	Agent List > Agent Dashboard	Create and Configure an Agent
Build and maintain data applications	Agent Console > Data Applications	Create Dashboard
Process user feedback and iterate	Agent Dashboard Top > Analyst Feedback Inbox	Feedback Loop
Improve Answer Accuracy	—	Agent Accuracy Tuning

Data analyst

Primary Tasks	Entry	Documentation
Discover and Subscribe to Agents	Agent Market	Agent Market and Subscription
Count the dialogues and verify the results	My Agent > Analyst Console	Start conversation analysis
Upload File and Ask Question	Analyst Workspace > File Upload	File Upload Analysis
Use Published Application	My Agent > Application View	Report Overview
Export and share results	Application Workspace	Data Export
Increase the hit rate of your questions	—	Recommended way to ask questions

System administrator and operations

Primary Tasks	Entry	Documentation
Platform initialization completed	System Management	Deploy by using a tool
Maintain identities and login channels	System Management > User and Login Integration	User and Authentication Management
Connect the model and bind it to the business module	System Management > Models	Connect to Model Service Provider
Maintain License and Capacity	System Management > License	License Activation and Capacity Management
Monitor the running status and locate issues	Monitor, Audit, Task Center, System Logs	Monitoring Center and Audit Logs

To be confirmed

Verification of consistency between feature names and actual menu names (The feature names in this table are derived from the description in the source manual; interface text has not been compared item by item.)
The minimum permission items corresponding to each feature must be aligned with the role permission list and then added to this table.