Verified project record

pingcap/tidb

TiDB is an open-source, cloud-native distributed SQL database that provides horizontal scalability, strong consistency, MySQL compatibility, and Hybrid Transactional and Analytical Processing (HTAP) capabilities. It addresses the problem of data silos and infrastructure ceilings for agentic workloads by combining transactions, analytics, and native vector search.

Project overview

It combines row-based and columnar storage engines alongside native vector search within a single distributed SQL database, using a two-phase commit and the Raft consensus protocol to maintain ACID compliance.

Project type: Vector Database · Data Processing · Infrastructure
Use cases: Knowledge Q&A
Deployment: Refer to project documentation
License: Apache-2.0

Best for

Developers, data teams, and enterprise teams who require a horizontally scalable, MySQL-compatible distributed SQL database for agentic workloads combining transactional, analytical, and vector search requirements.

Key capabilities

Uses a two-phase commit protocol to ensure ACID compliance and strong consistency across multiple nodes.
Scales horizontally by adding nodes or vertically by increasing resources without downtime, using separated computing and storage layers.
Built-in Raft consensus protocol ensures reliability, automated failover, and strong consistency using multiple data replicas.
Coordinates query execution across TiKV row-based and TiFlash columnar storage engines for optimized transactional and analytical performance.
Compatible with MySQL 8.0, allowing migration of applications without code changes or with minimal modifications.
Provides native vector search capabilities suitable for AI and agentic workloads.

Limitations and risks

May have minimal incompatibilities requiring code modifications during MySQL migration.

Getting started

Start a cluster using the local playground, Kubernetes, or TiDB Cloud. Afterward, learn the SQL syntax and connect a MySQL driver or ORM to build an application. The setup difficulty is rated medium.

Alternatives and comparisons

A vector similarity search engine supporting distributed deployment and multi-vector search, intended for applications focused strictly on vector storage and semantic matching rather than distributed SQL or HTAP workloads.
A graph database combining graph traversal, vector search, and temporal MVCC for systems requiring Neo4j-compatible Bolt/Cypher interfaces rather than MySQL-compatible SQL.
A single-file memory layer packaging data and embeddings without a database server, applicable for persistent AI memory needs that do not require distributed ACID SQL transactions.

Project comparisons

Evidence and sources

GitHub project description: TiDB is built for agentic workloads that grow unpredictably, with ACID guarantees and native support for transactions, analytics, and vector search. No data silos. No noisy neighb…
README: TiDB (/’taɪdiːbi:/, "Ti" stands for Titanium) is an open-source, cloud-native, distributed SQL database designed for high availability, horizontal and vertical scalability, strong…
README: TiDB can be deployed in public clouds, on-premises, or natively in Kubernetes.
Release: TiDB v8.5.7
README: TiDB can be easily deployed in a self-managed Kubernetes environment or Kubernetes services on public clouds using TiDB Operator.

Reviewed at: 2026-07-17T09:21:53Z

Open original GitHub project