Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Honeydew

honeydew.ai

  • Free trial
  • Web
  • API
6.8 Club score Visit honeydew.ai

What Honeydew is

Honeydew is a paid web and API semantic layer that brings business logic and context together for enterprise data, supporting AI interactions and business-intelligence queries. It models entities, metrics, and relationships from tables and business knowledge, then generates SQL for data consumers. Slack users and agents connected through MCP can receive answers based on shared definitions with access rules applied. Applications can call SQL and GraphQL APIs, with those requests compiled under built-in access rules. BI tools use the shared semantic model, governance, and performance optimizations while operating on a data warehouse. Documented warehouse accounts include customer-owned Snowflake, Databricks, and Google BigQuery. Integrations include Tableau, Power BI, Looker, GitHub, GitLab, Bitbucket, and Azure DevOps. Honeydew identifies itself as SOC 2 Type II compliant and says connections use TLS 1.2 or later. Customer metadata is stored in a private Git repository owned by the customer. Pricing starts at $20/user/month with a 14-day trial.

Who it suits

Honeydew suits data engineers, analysts, and AI builders contributing to a shared semantic model. It also fits organizations connecting BI tools, applications, and agents to governed warehouse data.

Good to know, and what to watch

  • Shared semantic model for AI and BI
  • SQL and GraphQL APIs include access rules
  • Supports governed metrics and row-level security
  • Integrates with major BI and development tools
  • 14-day trial includes modeling guidance
  • No free plan
  • Platform fees apply on Lite and Standard
  • Enterprise pricing is custom
  • Enterprise may include query-limited users

Enterprise

—

Lite

$20/mo

Standard

$30/mo

14-day trial

The verdict

Honeydew centralizes definitions, governance, and access controls for BI tools, applications, Slack users, and MCP agents. Its warehouse-based processing and shared workspace support coordinated data work, while pricing combines per-user charges with platform fees on listed plans.

What the maker publishes

AI interfaces
People using Slack and agents working over MCP can get answers compiled from shared definitions with access rules applied. source
APIs
Agents connect over MCP, while applications can call SQL and GraphQL APIs; requests compile with access rules built in. source
BI support
Honeydew lets BI tools use a shared semantic model, governance, and performance optimizations while operating on the data warehouse. source
Collaboration
Data engineers, analysts, and AI builders can contribute to the same semantic model in a shared workspace. source
Integrations
Documented integrations include Tableau, Power BI, Looker, GitHub, GitLab, Bitbucket, and Azure DevOps. source
Metadata storage
Customer metadata managed by Honeydew is stored in a private Git repository owned by the customer. source
Ongoing support
All tiers include product support during business hours with response SLAs based on issue severity. source
Product
Honeydew unifies business logic and context into a semantic layer for enterprise data, intended to support consistent AI interactions and BI queries. source
Query limits
The pricing FAQ says there are no query limits, though Enterprise contracts may include query-limited users. source
Security
Honeydew says application connections use TLS 1.2 or later and identifies itself as SOC 2 Type II compliant. source
Semantic modeling
It models entities, metrics, and relationships from tables and business knowledge, and generates SQL for data consumers. source
Trial support
The 14-day trial includes system access, support through a dedicated Slack or Teams channel, a training session, and up to four hours of data modeling guidance. source
Warehouses
Security documentation names Snowflake, Databricks, and Google BigQuery as customer-owned warehouse accounts where data processing occurs. source

Checked against the maker's pages on 4 Oct 2026.