Product Updates
AI
- Added support for Anthropic Claude Sonnet 5 LLM
- Added support for Anthropic Claude Opus 5 LLM
- Added support for OpenAI GPT-5.6 LLM
Integrations
- The Honeydew MCP server now includes
documentation access tools —
search_docsandquery_docs_filesystem— so a separate documentation MCP server is no longer needed. - Added Snowflake Intelligence MCP setup — configure Honeydew as an external MCP server in Snowflake to make Honeydew tools available to Snowflake Intelligence and Cortex Agents.
- Added Amazon Quick API integration — connect Honeydew to the Amazon Quick API to publish domains as datasets.
Security
- Added Windows Authentication (Kerberos) (Beta) for the XMLA endpoint — users on domain-joined Windows machines connect from Excel with their Windows login, without a password prompt or API key.
Governance
- Added a CI/CD workflow section — validate semantic-layer changes before merge and automate deployment from any CI/CD system.
- Added a GitHub Action,
Honeydew Validate Workspace
(
honeydew-ai/validate-workspace-action), available on the GitHub Marketplace. It validates a workspace on every pull request, so broken changes are blocked before reachingprod.
Interfaces
- Added the XMLA Interface for Excel (Beta) — Excel Pivot Tables connect to Honeydew natively, without requiring a Power BI license. Each domain is exposed as an OLAP cube, with entities as dimensions, attributes as hierarchies, and metrics as measures. See Excel for setup.
Integrations
- Added Sigma data model publishing (Beta) — publish a Honeydew domain to Sigma as a data model and build workbooks on the semantic layer.
AI
- Added agent routing to the Slack app — leave the Agent field blank to let Honeydew automatically route each question to the most relevant agent based on its description, AI context, and sample questions.
Integrations
- Coding agent plugins are available for Claude Code CLI, Claude Desktop, Cortex Code CLI, Cursor, GitHub Copilot CLI, Gemini CLI, Codex CLI, and Codex — installing the plugin sets up the Honeydew MCP servers automatically.
- Added Databricks BI integration — query the Honeydew semantic layer from Databricks dashboards via JDBC using the Trino protocol.
AI
- Added support for configuring the LLM model per agent via the
modelfield in the agent configuration. This overrides the workspace default model for deep analysis sessions started by that agent, allowing data teams to experiment with new model releases — including the latest Anthropic Claude Sonnet, Anthropic Claude Opus, OpenAI GPT, and Google Gemini Pro models — as they become available, subject to support by the workspace’s LLM provider.
Semantic Engine
- Released Domain Hierarchy (Beta), allowing domains to extend one or more parent domains and inherit their entities, filters, parameters, tags, and metadata
Migration
- Added support for bootstrapping from SAP Business Objects
AI
- Released Agents as the entry point for Deep Analysis. Agents define which domain and context the AI can access, and serve as the governance boundary for AI users.
- Released Context Management — a knowledge and memory layer for providing AI with organizational knowledge, analytical methodologies, and historical context. Supports inline content and external sources via MCP (Notion, Confluence).
AI
- Added MCP tools for semantic layer exploration and modeling
- Released the Claude Code plugins marketplace with skills for semantic modeling and data analysis
- Added support for Anthropic Claude Sonnet 4.6 LLM
- Added support for Anthropic Claude Opus 4.6 LLM
Migration
- Added support for bootstrapping from MicroStrategy
Data Sources
- Added support for Databricks as a data source
- Added support for Google BigQuery as a data source
AI
- Added support for Databricks Mosaic as an LLM Provider
- Added support for Google Vertex AI as an LLM Provider
- Added support for Gemini 3 Pro LLM model via Google Vertex AI, Snowflake Cortex and Databricks Mosaic
- Added support for Anthropic Claude Opus 4.5 LLM
UI
- Added new page for managing semantic context across all schema fields, enabling AI descriptions and sample value loading
AI
- Released Deep Analysis — a multi-step agentic workflow for investigative questions. Breaks complex questions into steps, executes targeted queries at each step, generates charts, and supports follow-up questions within the same conversation thread.
BI Integrations
- Added support for Tableau hierarchies through Tableau-specific metadata
- Added support for Power BI hierarchies through Power BI-specific metadata
- Added support for Power BI Row-Level Security (RLS)
AI
- Released MCP Server API for programmatic access to AI Analyst capabilities
- Added support for Anthropic as an LLM provider, including prompt caching
- Added prompt caching for Anthropic Claude models in Snowflake Cortex, reducing latency and cost
- Added support for Snowflake SEARCH function in AI-generated SQL queries, enabling more efficient full text search operations
BI Integrations
- Added support for Power BI Semantic Model Chaining
- Added support for 2-level folders in Looker metadata sync, using LookML
view_labelandgroup_label. For more details, see Looker metadata sync documentation.
Snowflake Integration
- Added support for deploying dynamic datasets as Snowflake Interactive Tables
- Added support for clustering expressions for deploying dynamic datasets as Snowflake Tables
AI
- Added support for Anthropic Claude 4.5 Sonnet LLM in Snowflake Cortex
- Added support for Snowflake AI_SUMMARIZE_AGG and AI_AGG AI functions in AI-generated SQL queries, enabling advanced summarization and aggregation capabilities on large text datasets
Honeydew Studio
- Added preview queries cancellation option
- Added query history pagination
Semantic Engine
Partial metrics - added support for removing all fields of an entity usingGROUP BY (NOT entity.*) syntax.AI / Security
Enhanced Slack App authentication with OAuth support, allowing individual users to authenticate with their own Honeydew accounts, enabling a better security model and enabling the use of personal Snowflake credentials when Snowflake OAuth authentication method is configured.Semantic Engine
- Added support for Conditional Filtering
- Improved support for Multi-Grain Tables
AI
Can answer complex data questions that require multi-step calculations (for example, cohort analysis).API
Released public GraphQL API, for programmatic access to the Semantic Layer and metadataBI Integrations
Security
Added support for IP Access ControlAI
- Answers can create calculated attributes as part of the response (binning, strings and dates manipulations, and more).
- Answers can create table calculations as part of the response (ratios, growth, percentages, % out of total, and more).
Integrations
Support for import from Snowflake Semantic Views to the Semantic Layer.BI Integrations
- Support for ThoughtSpot metadata sync
- Added ThoughtSpot-specific metadata
Security
Support for per-user API keys, enabling individual user identification within integrations. In combination with per-user Snowflake credentials this allows end-to-end user-level data access from any tool or API.AI
- Added support for Anthropic Claude 3.7 Sonnet and Claude 4 Sonnet LLM (using Snowflake Cortex AI)
- Added AI questions history and monitoring workflow in Honeydew Studio
- Reduced API calls latency by switching to Snowflake Cortex REST API
BI Integrations
Create new Honeydew data sources for Tableau directly from Honeydew UI.Git Integrations
Support for Azure DevOpsSecurity
- Support for OAuth authentication with Snowflake
- Support for OAuth authentication with Tableau
Onboarding
Added AI-assisted semantic model generation experience, to help users get started with the Semantic LayerBI Integrations
Access to generated LookML and Lightdash models through the Snowflake Native App API for programmatic integrationSecurity
- SSO with Microsoft Entra ID
- Support for PAT (programmatic access token) authentication with Snowflake
Semantic Engine
- Added support for custom relationship expressions, supporting complex join conditions and slowly changing dimensions
BI Integrations
- Added Lightdash-specific metadata
- Added support for Thoughtspot
Snowflake Integration
- Added Environments feature to customize Snowflake configuration per development branch
- Added support for multiple Snowflake connections or accounts
Performance and Stability
- Improved performance in high concurrency and high load environments
- Added support for long term query history
AI
- Add batch editing mode for AI metadata
- Detect modeling gaps and improve reliability with AI-as-a-Judge automatic validation
- Released a Teams App AI integration
Semantic Engine
- Improved support for Lightdash queries
- Improved support for Domo queries
- Improved support for multi-fact scenarios
Security and Compliance
- Added Honeydew Trust Center
BI Integrations
- Added support for Tableau conditional and top K filters
Performance
- Extended Aggregate Awareness support for shared dimensions
Modeling Experience
- Improved the performance for large entities (with over 1,000 fields)
- Improved the performance for large schemas (with over 100 entities)
AI
- User experience enhancements to Slack App such as Explain and Show SQL buttons
- Added support for AI-specific metadata in domains and fields, including sample values
AI
- Released a Slack App for AI self-service
- Added support for Anthropic Claude 3.5 Sonnet LLM (using Snowflake Cortex AI)
BI Integrations
Support for DomoSemantic Engine
Support for multi-fact schemas.BI Integrations
Direct publishing of Semantic Model to Power BI Service (use this manual to configure)Git Integrations
Governance
Support for Domain-level Role and Warehouse configurationAI
AI questions support in PlaygroundSemantic Engine
Extended support for conformed dimensionsBI Integrations
- Support for Excel Pivot Table
- Support for Tableau Server
Performance
Released aggregate awareness featureSemantic Engine
Released Cross-Filtering support for semantic modelsBI Integrations
- Support for Power BI-specific metadata sync
- Support for Tableau BI-specific metadata sync
- Support for Preset and Superset
- Support for Amazon Quick
AI
Released a Streamlit app for AI self-service, find more details hereSemantic Engine
Beta release of time aware metrics support for semantic modelsBI Integrations
- Support for Power BI
- Support for Hex
- Support for Mode Analytics
Performance
- Major performance improvements for supporting large and complex BI queries
- Query History page for monitoring query performance