TDengine IDMP 2.0.3 is officially available. This release introduces major updates across three areas: process recipe management and batch analysis, AI expert assistants, and Canvas visualization. It also includes improvements to access control, license management, system stability, and security.
Release Highlights
- Process Recipe Management and Batch Analysis Workbench: Organize production processes into reusable recipes and compare multiple batches by process stage.
- AI Expert Mode: Access specialized AI assistants, Deep Thinking Mode, and scheduled AI tasks.
- Canvas and Visualization: Simplify graphics configuration, inspect bound attributes, and improve dashboard interactions.
- Stability and Security: Expanded data connectivity, AI safeguards, and system fixes.
Process Recipe Management and Batch Analysis Workbench Are Here
Organize Batch Events with Reusable Process Recipes
Batch analysis is widely used in both discrete manufacturing and process industries.
Batch production typically involves multiple levels, including production orders, batches, equipment units, process stages, and individual operations. When product quality differs between batches, engineers often need to search through large volumes of historical data and perform manual comparisons to understand why.
IDMP 2.0.3 introduces a new Process Recipe Management module that allows users to configure batch hierarchies and associated events within the system. The same recipe can be reused across production lines.
Once a process recipe is configured, users can import the corresponding batch events and their associated attributes at different levels.
The event list also supports grouping by process recipe, with a Gantt chart displaying event durations and related information for different batches.
Batch events grouped by process recipe, with a Gantt chart showing event durations and related information
Compare Batches Across Process Stages
IDMP 2.0.3 also introduces a dedicated Batch Analysis Workbench.
Using the process stages defined in a recipe, engineers can compare multiple batches associated with the same recipe.
The workbench automatically aligns batches by the start time of each corresponding process stage, making it easier to compare process parameters within the same stage.
Process stages can be broken down into more detailed levels, and parameter charts can be configured as needed. Consistent color coding makes the charts easier to read, while clicking an individual batch highlights its corresponding data.
Analysis results can also be saved as panels and reopened with a single click.
These capabilities significantly reduce the effort required to compare production batches and investigate process differences.
The Batch Analysis Workbench automatically aligns batches by process stage and supports flexible comparisons of process parameters
AI Assistants Become More Specialized and Controllable
Expert Mode
IDMP 2.0.3 introduces Expert Mode for AI assistants.
Users can quickly access built-in domain-specific experts and predefined skills. Each expert has a defined role and scope of skills, allowing AI assistance to be more focused on particular tasks.
Deep Thinking Mode
The new Deep Thinking Mode displays the model’s reasoning output in a separate, expandable section.
Users can review the reasoning details behind a response. If the selected model does not support this capability, the system automatically falls back to its standard mode without interrupting normal use.
AI Scheduled Tasks
This release also introduces AI Scheduled Tasks for recurring analytical work.
Users can describe a task in natural language, and the system will execute it according to the configured schedule, returning the results to the original conversation.
Scheduled tasks support pausing, resuming, and retrying failed executions.
Users can access Expert Mode and Deep Thinking directly from the IDMP AI Chat interface
Better Security and Traceability
Built-in analytical skills can now retrieve knowledge graph information to provide supporting evidence, including entities, relationships, and confidence scores.
Reports can also be exported for further review.
For high-risk operations, IDMP adds an optional confirmation step before execution.
Content safety filtering has also been extended to user questions, tool parameters, and retrieved content, providing additional safeguards for industrial AI applications.
Canvas and Visualization Get More Practical
Canvas is an important tool for creating process visualizations and equipment monitoring interfaces.
Previously, creating a process diagram could require repeatedly adjusting the styling of graphical elements. When another engineer took over the work, configurations often had to be reviewed or adjusted again.
IDMP 2.0.3 introduces a Format Painter, allowing users to copy and reuse the styling of existing graphical elements with a single action.
New default configurations provide consistent settings for creating process diagrams, making it easier for multiple engineers to work with the same visualization standards.
Graphical elements can now display their bound attribute paths and current values when users hover over them. This makes it possible to inspect data bindings without repeatedly switching between the Canvas and attribute configuration pages.
Attribute value configuration has also been redesigned, and users can now add events directly to graphical elements on the Canvas.
Panel and Dashboard Improvements
Panels and dashboards have also received several updates.
Users can personalize appearance settings, with changes saved locally in their browsers so that individual preferences do not affect other users.
Image and PDF export options have been combined into a single entry point.
Trend charts in the Analysis Workbench also support box zoom and other interaction improvements.
Integration with CesiumJS for 3D Industrial Visualization
Beyond the Canvas improvements introduced in this release, TDengine IDMP can also be integrated with CesiumJS to build interactive 3D industrial visualization applications.
By combining 3D environments with industrial data from TDengine, users can explore equipment locations, operating conditions, and associated data within a visual representation of their facilities.
For example, TDengine demonstrations have shown how wind turbines and cement production lines can be represented in interactive 3D environments, with operating data and monitoring views linked to individual assets.
This integration makes it possible to explore industrial data in a spatial context alongside traditional charts and dashboards.
Interactive 3D industrial visualization using CesiumJS and TDengin
Note: CesiumJS integration is an existing capability demonstrated alongside this release, rather than a feature newly introduced in IDMP 2.0.3.
More Updates and Fixes
In addition to these major enhancements, IDMP 2.0.3 includes several other updates:
- Real-Time Analysis Insights: Real-Time Analysis Interpretation has been upgraded to Real-Time Analysis Insights, generating validated decision-support reports. When validation fails, the system provides the available evidence and relevant limitations.
- SQL Server Support: Federated query capabilities now support Microsoft SQL Server.
- Parameterized External SQL: External SQL calls used by tag attributes now support attribute parameters.
- OpenAI Codex Integration: The OpenAI Codex plugin for IDMP is now available.
- Model Management: Model management now supports non-time-series datasets, along with improvements to anomaly detection and classification modeling.
Bug Fixes and Stability Improvements
This release also addresses several issues:
- Knowledge graph extraction from documents now adapts to the remaining model context when approaching token limits, reducing extraction failures caused by truncation.
- LLM API keys are now obtained dynamically instead of being hardcoded, improving AI security.
- Fixed an issue where stream processing notifications could not reach IDMP in Docker deployments.
- Upgraded Keras in TDgpt to address CVE-2026-1669.
- Improved count-window stream processing by automatically adding the required
WHEREconditions.
Upgrade to IDMP 2.0.3
We recommend that all IDMP users upgrade to version 2.0.3.
If you are upgrading from an earlier version, back up your data before proceeding.
For upgrade instructions or assistance, refer to the official IDMP 2.0.3 Release Notes or contact the TDengine technical support team.
Modernize Your Industrial Data Architecture Without Starting Over
Bring existing industrial data into a common context for modeling, relationships, analytics, and AI, without making full data migration the first step.
