llms.txt
Compare with other sites' llms.txt files →# Bytebase
> Bytebase is a database governance platform: one gateway through which engineers and AI agents change, query, and audit databases. It holds the database credentials, applies SQL review, approval, just-in-time access, dynamic data masking, and an audit log to every change and query, across PostgreSQL, MySQL, SQL Server, Oracle, MongoDB, Snowflake, ClickHouse, and 20+ engines. Open-source; self-hosted or Bytebase Cloud.
## When to use Bytebase
- **Engineers hold direct production credentials and nobody can prove who ran what.** Route every query and export through one gated, audited path; Bytebase holds the credential.
- **SQL is run by hand against production with no review step.** Each change becomes an issue with automated SQL review, risk-tiered approval, rollout, and rollback.
- **An auditor or SOC 2 assessor is asking for the log.** One audit format across engines, tied to SSO identities, exportable over the API.
- **The fleet outgrew hand-managed grants.** Dozens to hundreds of instances across engines and clouds, governed from one place, with batch change across tenants.
- **Schema changes are not treated like code.** Review, approval, and audit around the migration files your ORM, Liquibase, or Flyway already produce.
- **Sensitive columns are visible to anyone who can query.** Dynamic data masking on the Bytebase read path, with unmask approval.
- **An AI agent needs to read or change a database.** It connects to Bytebase over MCP, not to the database, under its own identity, with masking, review, approval, and audit.
Bytebase is not a database, a secrets manager, a general privileged-access product, or a data catalog. Masking and query control cover reads through Bytebase, not the application's own connection. The full fit guide, including what it is not: /index.md.
## Start here
- [Bytebase](/index.md): what Bytebase is, who it is for, when to use it and when not to, what it does, supported databases, pricing, trust, and how an agent can act on it.
- [Database Governance for AI Agents](/database-governance-ai-agents.md): the identity, authorization, isolation, audit, and masking model for agents.
- [Database MCP Server](/database-mcp-server.md): the built-in MCP server, its workspace access policy, and client setup.
- [Pricing](/pricing.md): plan contents and which features are gated by tier.
- [Bytebase, compared](/compare.md): every head-to-head comparison with a migration tool, SQL client, or access platform.
## How an agent can act on Bytebase
- MCP server, built into every Bytebase instance at `<external_url>/mcp` (HTTP, OAuth sign-in). Sessions run under the signed-in identity, capped by a workspace access policy (Disabled, Read-only, Read-write). Docs: https://docs.bytebase.com/integrations/mcp
- REST API at https://api.bytebase.com, authenticated with a service account key as `Authorization: Bearer` or with workload identity (short-lived OIDC token from CI). Covers issues, plans, SQL review, permissions, and audit log export.
- Terraform provider for instances, projects, environments, roles, and policies: https://docs.bytebase.com/integrations/terraform/overview
- GitOps: migration files committed to GitHub, GitLab, Bitbucket, or Azure DevOps open a reviewed, approved rollout: https://docs.bytebase.com/gitops
## How to read this site as markdown
- Append `.md` to any marketing, blog, customer, or reference page URL: https://www.bytebase.com/pricing/ becomes https://www.bytebase.com/pricing.md. Every HTML page advertises its mirror with `<link rel="alternate" type="text/markdown">`.
- Every marketing, comparison, and customer page concatenated into one document: https://www.bytebase.com/llms-full.txt
- Documentation index: https://docs.bytebase.com/llms.txt
## Product
- [About Bytebase](/about.md): We're changing the way teams and AI agents work with databases
- [Bytebase for Enterprise](/enterprise.md): Standardize how teams and AI agents change, access, and audit databases across the enterprise.
- [Database Change Management](/database-change-management.md): Bytebase automates database change workflows with built-in review, rollback, and optional GitOps integration so teams work from the same process.
- [Database Audit Logging](/database-audit-logging.md): An audit trail of every database action tied to a real identity: SSO identity, full SQL, approvals, and masking decisions captured at the gateway — one format across PostgreSQL, MySQL, SQL Server, and Oracle, mapped to SOC 2, HIPAA, and PCI DSS. Bytebase records the workflow layer (actions routed through it); connections that bypass it need engine-native auditing as a backstop. Self-hosted, open-source.
- [Database as Code](/database-as-code.md): Run your database as code: schema and migrations in version control, automated SQL review, pull-request approvals, CI/CD deployment, and an audit trail. Self-hosted, open-source.
- [Database Change Automation](/automate-database-change.md): Automate database change end to end: a GitOps change pipeline, automated SQL review, risk-tiered approvals, rollback, and a complete audit trail. Self-hosted, open-source.
- [Database Governance for AI Agents](/database-governance-ai-agents.md): Govern what AI agents read and write in your database. In Bytebase an agent authenticates with its own identity — a service account or workload identity — that inherits the same governance as a human user — just-in-time access, masking, and audit on reads, plus SQL review, approval, and rollback on writes — over MCP, across PostgreSQL, MySQL, SQL Server, and Oracle. Not vanilla IAM (auth only) or PAM (credential/session layer). Self-hosted, open-source.
- [Database MCP Server](/database-mcp-server.md): The Bytebase MCP server: governed reads and changes for AI agents, built into Bytebase. The agent connects to Bytebase, not the database. A workspace access policy (Disabled, Read-only, Read-write; Read-only for new workspaces) caps every MCP session, each session runs under the signed-in user's permissions, sensitive columns are masked on the read path, writes go through SQL review and approval, and every query and change is audit-logged with an MCP tag. Self-hosted or Bytebase Cloud, open-source, PostgreSQL, MySQL, SQL Server, Oracle, and 20+ engines.
- [SQL Review](/sql-review.md): Catch unsafe SQL before it reaches production: 100+ automated review rules across linting, schema-aware semantic checks, and risk-based approval, wired into your CI/CD pipeline. Self-hosted, open-source.
- [Online Schema Migration for MySQL](/online-schema-migration.md): Run large MySQL schema changes without downtime. Bytebase picks Online DDL or gh-ost by table size, drives the shadow-and-cutover, and throttles on replica lag — all behind SQL review, approval, and audit. Self-hosted, open-source.
- [Dynamic Data Masking](/dynamic-data-masking.md): Mask sensitive columns at query time, not at rest: role-based policies, classification-driven rules, unmask approval, and a full audit trail — one policy across PostgreSQL, MySQL, SQL Server, Oracle, BigQuery, and Snowflake. Masking covers the human read path through Bytebase (SQL Editor, query results, exports), not the application's direct connection. Self-hosted, open-source.
- [SQL Editor](/sql-editor.md): Query production through one governed web SQL editor: shared connections, just-in-time access, dynamic data masking, and a full audit log. Open-source.
- [Just-in-Time Database Access](/just-in-time-database-access.md): Replace standing privileges with just-in-time database access: request, approve, and auto-revoke time-bound, least-privilege grants, fully audited. Open-source.
- [Database Access Control](/database-access-control.md): Grant least-privilege, just-in-time access through approval workflows, with dynamic data masking and a full audit trail. A zero-trust path to production data — no standing credentials, every query accountable.
- [Database Compliance](/database-compliance.md): From schema changes to data access, Bytebase records who did what, when, and why across every environment.
- [Bytebase Pricing](/pricing.md): Community (free), Pro ($20/user/month), and Enterprise (custom) plans.
- [Bytebase, compared](/compare.md): Bytebase compared, side by side: schema migration tools, SQL clients, and database access platforms.
- [Amazon RDS Pricing (AWS) — On-Demand & Reserved](/dbcost/rds-pricing.md): On-demand and reserved monthly pricing for every Amazon RDS (AWS) instance, across PostgreSQL, MySQL, and MariaDB, in every region.
- [Amazon Aurora Pricing (AWS) — On-Demand & Reserved](/dbcost/aurora-pricing.md): On-demand and reserved monthly pricing for every Amazon Aurora (AWS) instance, across Aurora PostgreSQL and Aurora MySQL, in every region.
- [Google Cloud SQL Pricing (GCP) — Enterprise & Enterprise Plus](/dbcost/cloudsql-pricing.md): On-demand monthly pricing for every Google Cloud SQL machine type, across Enterprise and Enterprise Plus, for MySQL and PostgreSQL, in every region.
- [Azure Database for PostgreSQL & MySQL Pricing — Flexible Server](/dbcost/azure-flexible-server-pricing.md): On-demand and reserved monthly pricing for every Azure Database for PostgreSQL and MySQL Flexible Server machine type, across Burstable, General Purpose, and Memory Optimized, in every region.
- [Google AlloyDB Pricing (GCP)](/dbcost/alloydb-pricing.md): On-demand monthly pricing for every Google AlloyDB machine shape — PostgreSQL-compatible, per vCPU and per GB, in every region.
- [Azure SQL Database Pricing — vCore, Every Service Tier](/dbcost/azure-sql-database-pricing.md): On-demand and reserved monthly pricing for every Azure SQL Database vCore configuration, across General Purpose, Business Critical, and Hyperscale, in every region.
- [PostgreSQL Data Masking — Dynamic Masking Without an Extension](/databases/postgres/data-masking.md): Role-based dynamic data masking for PostgreSQL. No extension, no superuser — one policy across community Postgres, RDS, Aurora, Cloud SQL, and AlloyDB, with approval workflow and audit.
- [PostgreSQL Schema Migration — Database CI/CD for Postgres](/databases/postgres/schema-migration.md): A Postgres schema change is one ALTER TABLE — instant, ungoverned, and hard to take back. Bytebase puts it on a CI/CD pipeline: every change reviewed, rolled out by stage, approved, and audited across community Postgres, RDS, Aurora, Cloud SQL, and AlloyDB.
- [MySQL Data Masking — One Policy Across Every Edition and Fork](/databases/mysql/data-masking.md): Role-based dynamic data masking for MySQL. No Enterprise license, no view inventory, no proxy — one policy across Community, Enterprise, Percona, MariaDB, RDS, and Aurora, with approval workflow and audit.
- [MySQL Schema Migration — Database CI/CD for MySQL](/databases/mysql/schema-migration.md): MySQL has no transactional DDL — a multi-statement migration that fails leaves the table half-applied, with no rollback. Bytebase puts MySQL change on a CI/CD pipeline: reviewed, run online through gh-ost when an ALTER would lock, staged, approved, and audited across Community, Percona, MariaDB, RDS, and Aurora.
- [SQL Server Data Masking — Beyond Native DDM's Display Filter](/databases/sql-server/data-masking.md): SQL Server has Dynamic Data Masking, but Microsoft calls it a display filter, not a security boundary — WHERE-predicate inference leaks the real value. Bytebase governs the query path with role-based masking, approval workflow, and audit across SQL Server, Azure SQL, and RDS.
- [SQL Server Schema Migration — Database CI/CD for SQL Server](/databases/sql-server/schema-migration.md): On SQL Server the deploy script is often generated — a DACPAC diff or an SSMS one-off that runs before anyone reads it. Bytebase puts SQL Server change on a CI/CD pipeline: every change reviewed, staged, approved, and audited across SQL Server, Azure SQL, Managed Instance, and RDS — so a generated drop-and-add never reaches production unread.
- [Oracle Data Masking — Past Data Redaction's License and Leaks](/databases/oracle/data-masking.md): Oracle Data Redaction is a paid Advanced Security add-on on Enterprise Edition — and a WHERE predicate still leaks the value while SYSDBA and EXEMPT holders read cleartext. Bytebase governs the query path with role-based masking, approval, and audit across every Oracle deployment, no license required.
- [Oracle Schema Migration — Database CI/CD for Oracle](/databases/oracle/schema-migration.md): An Oracle schema is a web of dependencies — alter one table and the packages, views, and triggers that reference it go INVALID. Bytebase puts Oracle change on a CI/CD pipeline: reviewed, staged so the invalidations surface first, approved, and audited across Enterprise, Standard, Autonomous, Exadata, and RDS.
- [PostgreSQL MCP Server — Governed Agent Access to Postgres](/databases/postgres/mcp-server.md): A governed MCP server for PostgreSQL. Agents connect through Bytebase, not a connection string — per-agent identity, masking on the read path, writes reviewed as change issues, and every action audited.
- [MySQL MCP Server — Governed Agent Access to MySQL](/databases/mysql/mcp-server.md): A governed MCP server for MySQL. Agents connect through Bytebase, not a DSN — per-agent identity, masking on the read path, writes reviewed as change issues, and every action audited across Community, Percona, MariaDB, RDS, and Aurora.
- [SQL Server MCP Server — Governed Agent Access to SQL Server](/databases/sql-server/mcp-server.md): A governed MCP server for SQL Server. Agents connect through Bytebase, not a connection string — per-agent identity, masking that sysadmin and native DDM can't bypass, writes reviewed as change issues, and every action audited across SQL Server, Azure SQL, and RDS.
- [Oracle MCP Server — Governed Agent Access to Oracle](/databases/oracle/mcp-server.md): A governed MCP server for Oracle. Agents connect through Bytebase, not a connection string — per-agent identity, masking with no Advanced Security license that SYSDBA can't bypass, writes reviewed as change issues, and every action audited across Enterprise, Standard, Autonomous, and RDS.
- [Bytebase vs. Atlas](/compare/bytebase-vs-atlas.md): Compare Bytebase and Atlas for database schema migration: declarative vs. issue-based workflows, GitOps, team collaboration, and enterprise governance.
- [Bytebase vs. CloudBeaver](/compare/bytebase-vs-cloudbeaver.md): Compare Bytebase and CloudBeaver for web-based database management: approval workflows, data masking, audit logs, and centralized access control.
- [Bytebase vs. CyberArk](/compare/bytebase-vs-cyberark.md): Already run CyberArk? Compare it with Bytebase, the database-specific PAM: schema change + SQL review + dynamic masking + JIT query access in one auditable platform.
- [Bytebase vs. DataGrip](/compare/bytebase-vs-datagrip.md): Compare Bytebase and DataGrip for SQL workflows: individual IDE productivity vs. team-based access control, data masking, audit logs, and governance.
- [Bytebase vs. DBeaver](/compare/bytebase-vs-dbeaver.md): Compare Bytebase and DBeaver for SQL workflows: client-side desktop productivity vs. centralized access control, data masking, and team collaboration.
- [Bytebase vs. DbUp](/compare/bytebase-vs-dbup.md): Compare Bytebase and DbUp for database change management: review, approval, batch rollout, access control, and audit around the SQL scripts a .NET runner applies.
- [Bytebase vs. Delphix](/compare/bytebase-vs-delphix.md): Compare Bytebase and Delphix for data masking: masked copies for non-production versus dynamic masking on live production queries, with schema change, SQL review, and access control in one platform.
- [Bytebase vs. Flyway](/compare/bytebase-vs-flyway.md): Compare Bytebase and Flyway for database schema migration: GUI-based collaboration, GitOps, batch change, and enterprise governance across teams.
- [Bytebase vs. Jira](/compare/bytebase-vs-jira.md): Compare Bytebase and Jira for database change management: purpose-built issue-based review, approvals, GitOps, and audit-ready schema delivery.
- [Bytebase vs. Liquibase](/compare/bytebase-vs-liquibase.md): Compare Bytebase and Liquibase for database schema migration: GUI-based collaboration, GitOps, batch change, governance, and enterprise-grade review workflows.
- [Bytebase vs. Metabase](/compare/bytebase-vs-metabase.md): Compare Bytebase and Metabase — database development governance vs. BI analytics — and see how each tool fits a different seat in the data stack.
- [Bytebase vs. Navicat](/compare/bytebase-vs-navicat.md): Compare Bytebase and Navicat for SQL workflows: single-user desktop client vs. centralized credentials, data masking, review workflows, and audit logs.
- [Bytebase vs. PopSQL](/compare/bytebase-vs-popsql.md): Compare Bytebase and PopSQL before the September 1, 2026 shutdown: shared query library, access control, data masking, and audit logging for migrating teams.
- [Bytebase vs. SchemaHero](/compare/bytebase-vs-schemahero.md): Compare Bytebase and SchemaHero for database schema management: review, approval, access control, and audit around every schema change, on or off Kubernetes.
- [Bytebase vs. StrongDM](/compare/bytebase-vs-strongdm.md): Compare Bytebase and StrongDM for database access control: schema change + SQL review + dynamic masking + JIT query access in one auditable platform.
- [Bytebase vs. Teleport](/compare/bytebase-vs-teleport.md): Compare Bytebase and Teleport for database access control and JIT: per-query audit, object-level permissions, SQL review, dynamic masking, and pricing.
## Cloud Database Pricing
- [Amazon RDS instance pricing](/dbcost/rds/instance.md): Per-instance Amazon RDS pricing — specs and monthly cost across regions and engines.
- [Amazon RDS instance comparisons](/dbcost/rds/compare.md): Compare Amazon RDS instances by size, side by side.
- [Amazon Aurora instance pricing](/dbcost/aurora/instance.md): Per-instance Amazon Aurora pricing — specs and monthly cost across regions and engines.
- [Amazon Aurora instance comparisons](/dbcost/aurora/compare.md): Compare Amazon Aurora instances by size, side by side.
- [Google Cloud SQL instance pricing](/dbcost/cloudsql/instance.md): Per-instance Google Cloud SQL pricing — specs and monthly cost across regions and engines.
- [Google Cloud SQL instance comparisons](/dbcost/cloudsql/compare.md): Compare Google Cloud SQL instances by size, side by side.
- [Google AlloyDB instance pricing](/dbcost/alloydb/instance.md): Per-instance Google AlloyDB pricing — specs and monthly cost across regions and engines.
- [Azure Database for PostgreSQL / MySQL Flexible Server instance pricing](/dbcost/azure-flexible/instance.md): Per-instance Azure Database for PostgreSQL / MySQL Flexible Server pricing — specs and monthly cost across regions and engines.
- [Azure Database for PostgreSQL / MySQL Flexible Server instance comparisons](/dbcost/azure-flexible/compare.md): Compare Azure Database for PostgreSQL / MySQL Flexible Server instances by size, side by side.
- [Azure SQL Database instance pricing](/dbcost/azure-sql/instance.md): Per-instance Azure SQL Database pricing — specs and monthly cost across regions and engines.
## Brand & Design
- [Brand Guidelines](/brand.md): The full Bytebase brand system — strategy, verbal identity, logo, brand device, color, typography, iconography, illustration.
- [Design Tokens](/design.md): Machine-readable brand color and typography tokens in the design.md format, the source of truth mirrored by the Brand Guidelines.
## Customer Case Studies
- [PropertyGuru — How PropertyGuru Runs Database Governance as Code with Bytebase](/customers/propertyguru.md): How PropertyGuru automates database changes and access by managing projects, permissions, and policies as code with Bytebase.
- [Lucid Software — How Lucid Gives Engineers Safe, Self-Service Access to Production Databases with Bytebase](/customers/lucid.md): How Lucid Software replaced an ops-gated, m