Bytebase

Bytebase is an open-source database governance platform that sits as a control plane between users — both human engineers and AI agents — and an organization's databases. It consolidates functions that teams otherwise assemble from migration scripts, standalone SQL clients and ticketing systems, so that database changes and queries are reviewed, controlled and recorded in one place.

The platform is organized around three areas. Change management provides a GUI workflow to plan, request, review, approve, deploy and roll back schema and data changes, GitOps integration with GitHub and GitLab for database-as-code, and an SQL review engine with more than 200 lint rules. Access control offers project- and workspace-level RBAC, just-in-time time-boxed access grants with automatic revocation, dynamic column-level data masking applied at query time, and IAM policy enforcement. Compliance features include immutable audit logging of database activity, policy-as-code via a Terraform provider and API, and data classification for tagging sensitive data. AI-oriented capabilities include a Model Context Protocol server that gives each agent its own identity while routing its changes and queries through the same masking, review and audit controls, plus text-to-SQL assistance in the SQL editor and a built-in assistant that executes workflows from plain-language requests.

Working concepts in the product include projects (containers grouping databases and related tasks), instances (database deployments registered with the platform), environments (including designated production environments), issues used to track changes and requests, a SQL editor with syntax highlighting, query history and tabular results, and slow query analysis built on MySQL slow query logs and PostgreSQL pg_stat_statements.

Founding story

The company states that Bytebase was built by former tech leads of Google Cloud SQL and Google's API infrastructure, motivated by the observation that while application code and infrastructure became programmable and version-controlled, database work was still handled through ad-hoc scripts, tickets and shared credentials, creating risks from manual changes, over-privileged access and audit gaps.

Business model

Bytebase is distributed as an open-source platform under a dual-license model (a general license plus a separate enterprise license) and is offered both as a self-hosted deployment and as a hosted cloud service. Self-hosted installation is available via Docker images and a Helm chart for Kubernetes, with paid/enterprise functionality layered on top of the freely available core.

Revenue comes from commercial offerings around the open-source core: a hosted cloud service and self-hosted enterprise deployments governed by a separate enterprise license.

Traction

The open-source repository has roughly 14.5k stars, 984 forks and 22,463 commits, and the company reports more than 5 million downloads worldwide, support for a large number of database types, and SOC 2 and HIPAA certification. Published case studies report Lucid Software reducing query turnaround from 5–20 minutes to under a minute via self-service production access, note.com implementing just-in-time database access while reviewing thousands of changes, and Longbridge Whale unifying tenant databases for schema changes.

Latest developments

Recent product direction centers on governing AI agent access to databases: an MCP server that assigns each agent its own identity and routes its activity through existing masking, review and audit policies; text-to-SQL generation and refinement in the SQL editor; and a page agent that guides or executes workflows from natural-language requests. In May 2026 the company published an announcement positioning Bytebase as the standard for database development.