Release Notes for Q1 2026

Calibo Accelerate now enables the creation of Agentic AI workflows. The platform integrates with CrewAI to orchestrate multiple AI agents that collaborate to perform complex tasks. Each agent is assigned a specific role and responsibility within the workflow.

Users can define the desired intent, based on which the platform automatically generates a workflow design. The design can be iteratively refined and reviewed. Once finalized, the platform generates the corresponding code for each agent. Users can then deploy the agents in a Kubernetes cluster.

Calibo Accelerate now supports integration with OpenAI and Google Gemini as Large Language Model (LLM) providers. After you configure the respective connections, you can create and manage model profiles to control access to specific LLMs within the platform.

You can associate one or more model profiles with a product and designate a default model profile for general use, along with task-specific model profiles for targeted operations or workflows. Model profiles provide a structured way to define the models available to a product and control how those models are used across different tasks and workflows.

This enhancement provides flexibility and control over how LLMs are accessed and utilized across the Calibo Accelerate platform.

Users with administrator access to LLMs can edit all model profiles associated with a given LLM. Access for users and teams is managed at the model profile level for each product, providing granular control over model usage.

You can also create custom roles and assign permissions specific to LLMs and their associated model profiles.

When deactivating a user, the Calibo Accelerate platform now displays information to help the user proceed with the deactivation:

• User’s associations with workflow templates and approval workflows. This visibility helps you identify and remove the user from relevant templates or workflows before proceeding with deactivation.

• A “Deactivated” tag is displayed alongside the user’s name across applicable screens of the platform like product, teams, users and so on. This is helpful while assigning users to teams or roles.

Calibo Accelerate now supports handling schema changes in source data for Unity Catalog DQ Processor jobs. When Auto-Evolve Target Schema is enabled and a schema change is detected, it is flagged as Updates Available in the Schema Mapping stepper.

Click Updates Available to review detected changes, including column additions and deletions. You can update the schema mapping, including modifying column names as needed.

Previously, DQ Processor reports could be exported only as consolidated PDF. This release enhances the download functionality to provide greater flexibility. You can now export individual DQ Processor widgets or download a consolidated report in CSV and Excel formats, based on your business needs.

When you export the consolidated report in Excel format, the report for each widget is downloaded in individual sheets of the Excel file.

Calibo Accelerate now supports profiling existing data in a Unity Catalog data lake. This capability is available at the platform level under Data Lab > Data Profiling.

You can create a data profiling job by selecting a Databricks cluster and a Unity Catalog data lake. The job can be run on demand or it can be scheduled, and profiling results can be viewed directly within the platform, without the need to create a pipeline in Data Pipeline Studio.

Calibo Accelerate now introduces the ability to create Retrieval-Augmented Generation (RAG) pipelines through Data Pipeline Studio. 

RAG enhances LLM responses by incorporating information from external knowledge sources, rather than relying solely on pre-trained data.

This enables organizations to leverage domain-specific data without retraining the underlying model, while ensuring that the responses reflect current and proprietary information.

In this release, RAG pipelines are built exclusively on Databricks Unity Catalog, leveraging its native data sources, embedding models, vector search and LLM offerings.

With RAG Pipeline Builder, you can create a RAG pipeline within a few clicks. You can:

• Ingest data from source formats such as .doc, .docx, .ppt, .pptx, .pdf

• Apply chunking strategies to split data into manageable segments

• Use embedding models to convert text into vector representations

• Store embeddings in a vector database index

• Configure a retriever strategy to fetch relevant context

• Generate responses using an LLM

Calibo Accelerate now supports creating data analytics pipelines that include multiple JupyterLab instances. During technology deployment, each JupyterLab instance must be assigned a unique context path to ensure independent access to each instance.

You cannot add another JupyterLab instance to an existing pipeline or feature. However existing pipelines with a single JupyterLab instance continue to function without any changes.

Note:

If JupyterLab is uninstalled and then reinstalled, existing pipelines that reference that JupyterLab deployment will fail.

Calibo Accelerate now supports query tagging and object tagging for integration, transformation, and data quality jobs with Snowflake as the target data lake. This enables improved governance, traceability, and cost tracking.

When configuring a Snowflake account in the Calibo Accelerate platform, the platform automatically adds two system-defined query tags, by default. The system tags are applicable at both the Snowflake instance and job levels. You can modify this setup by deleting the default tags and defining your own system query tags as key–value pairs.

While configuring Snowflake within a job, you can:

• Use system tags for tracking query execution across the instance and jobs

• Define custom (object) tags, which are applicable only at the job level.

These tags can be tailored to your business need to track query execution, attribute costs, and enhance observability.

Calibo Accelerate now allows you to define any branch as default within branch templates. The updated behavior applies as follows:

• Default Branch Configuration 

  You can now define any branch as the default branch within a branch template. 

• Behaviour of new and existing repositories

  • The updated branching strategy is automatically applied to all newly created repositories.

  • Existing repositories are not impacted by changes to the default branch or branch deletions.

  • Newly added branches are automatically created for existing repositories.

Calibo Accelerate now supports Traefik as an Ingress Controller for Kubernetes clusters.

This support is available for the following deployments:

• Calibo Accelerate Orchestrator Agent (AWS and Azure)

  Enables access to agent services through automatically configured ingress routes, ensuring secure and reliable communication.

• Technology and Applications (AWS EKS and Azure AKS)

  Applications deployed using Traefik automatically generate live URLs, enabling direct access without additional manual configuration.

This enhancement enables secure and scalable routing of external traffic, with built-in load balancing, and flexible routing capabilities.

Calibo Accelerate now extends the support for Snyk (Cloud) as a container image security scanning tool for Docker and Kubernetes deployment modes with Jenkins as a CI/CD tool.

You can configure Snyk centrally, select it at the stage level as the Container Image Scanning Tool, and enforce severity-based security gates that automatically fail CI pipelines when vulnerabilities exceed the defined thresholds.

Snyk scan results are visible across pipeline views, technology deployment details, and monitoring dashboards, with full audit logging and direct links to the associated Snyk projects.

Calibo Accelerate now extends support for serverless application deployments to AWS Lambda using AWS Serverless Application Model (SAM) with Jenkins CI/CD pipelines. You can configure deployment stages using the serverless mode, add Lambda-based technologies, and deploy functions using SAM and AWS CloudFormation.

With this support, you can configure all required AWS Lambda deployment options—such as the SAM template path, IAM roles, CloudFormation parameters, and artifact locations directly from the Calibo Accelerate UI.

You can also monitor deployment progress, view logs and metrics, and test Lambda functions from within the platform, without switching to the AWS console.

Calibo Accelerate now supports Helm 4 for application deployment in Kubernetes clusters. You can select a Helm 4 template to deploy and manage applications and platform components with greater consistency and simplified configuration.

You can use Helm 4 for Calibo Accelerate Orchestrator Agent deployment.

Support for New Technologies

Support for the following latest versions of technologies is now available on the Calibo Accelerate platform:

• Angular 21

• ASP.Net Core 10

• Nuxt 4

• NextJS 16

• NextJS 16 with Pnpm

• React 19 with TypeScript - Pnpm

• NestJS 11

• NestJS 11 with Pnpm

• Helm Template v4

• Java17 Spring Boot 4.0-Gradle

• Java17 Spring Boot 4.0-Maven

• Java17 with GraphQL Spring Boot 4.0-Gradle

• Java17 with GraphQL Spring Boot 4.0-Maven

• Java21 Spring Boot 4.0-Gradle

• Java21 Spring Boot 4.0-Maven

• Java21 with GraphQL Spring Boot 4.0-Gradle

• Java21 with GraphQL Spring Boot 4.0-Maven

To view all the supported tools and technologies, see Tools and Technologies Integrated with Calibo Accelerate Platform.

Deprecated Technologies

The following End-of-Life (EOL) versions of technologies are no longer supported in Calibo Accelerate:

• Nuxt-3.2.2

• NestJS 8

• NestJS 9

• NextJS 14

• React 16.0

• React 16.0 without TypeScript

• React 17.0.2 with TypeScript - Yarn

• React 17.0.2 without TypeScript - Yarn

• React 17.0.2 without TypeScript

• Amazon Elastic Kubernetes Service (EKS) 1.32