Release Date: Jun 16th, 2026

1. Overview

Version 2.20.0 delivers a major expansion of Brighthive's platform: MCP-based connectivity for platform agents, automated Bring-Your-Own-Warehouse ingestion, a full semantic view lifecycle for dbt, a configurable data quality rule library, enterprise Okta authentication, and longitudinal data health monitoring.

Highlights include:

• MCP Connectivity: The dbt agent and other platform tools are now reachable over the Model Context Protocol (MCP), with a live connectivity test and prompt playground in the web app.

• Automated BYOW Ingestion: Snowflake and Redshift catalogs now flow automatically into embeddings-based retrieval, scoped safely to production schemas.

• Semantic View Lifecycle: The dbt agent can now write semantic views back to Snowflake and ship them as governed, CI-checked pull requests to GitHub.

• Configurable Quality-Rule Library: 20+ ready-to-use data quality rule templates, with full create/edit/run support and pass-rate history tracking.

• Enterprise Auth: Added Okta federation for secure enterprise single sign-on.

• Longitudinal Monitoring: Introduced anomaly detection that tracks data health trends over time.

  
  

2. Key Enhancements and Fixes

MCP Integration

• Made the dbt agent, and other platform tools, reachable over MCP, with a dedicated connectivity endpoint and a live connectivity card in the web app showing available tools and a prompt playground.

Ingestion & BYOW

• Automated the ingestion pipeline for Bring-Your-Own-Warehouse customers: Snowflake and Redshift catalogs are automatically embedded and made retrievable, scoped to production schemas only.

• Improved webhook routing to correctly direct events to the right workspace.

dbt & Semantic Views

• The dbt agent can now write semantic views directly back to Snowflake.

• Semantic views can be shipped as governed pull requests to GitHub, with automatic CI compile checks.

• Added lineage tracking for semantic views, showing base tables and join relationships.

Data Quality

• Introduced a configurable quality-rule library with 20+ templates, full create/edit/run support, and pass-rate history tracking in the web app.

Notifications & Human-in-the-Loop

• Added Human-in-the-Loop input handling for dbt query-selection interrupts.

• Added monitoring and an on-call runbook for notification delivery.

Enterprise Auth & Monitoring

• Added Okta federation support for enterprise single sign-on.

• Introduced longitudinal anomaly detection to monitor data health trends over time.

3. Feedback

We value your feedback to continue improving the Brighthive platform. Please share your thoughts or report issues to support@brighthive.io.

About Brighthive 

Brighthive is an agentic data platform that functions as a data team in a box — a coordinated set of specialized AI agents (BrightAgent) that work the full data lifecycle around the clock: connecting to your sources, cleaning and validating data, enforcing governance and data contracts, and keeping pipelines self-healing and live. Underneath every workspace, Brighthive builds and maintains a living knowledge graph and enterprise data catalog — capturing not just schemas and metadata but the tribal knowledge about your business, use cases, and policies that normally lives only in people's heads. That context is what lets the platform, and every downstream tool, reason accurately about your data instead of guessing. The result: every team inside the enterprise gets to skip the grunt work and go straight to trustworthy, self-service insight — querying, visualizing, and analyzing data faster, on a foundation that's clean and governed by default. Brighthive connects to your existing data stack or stands one up from scratch, augmenting or replacing point tools without disrupting how your team already works.