TiDB Cloud Starter: Auto-Scaling for Dynamic Workloads

Trusted by Teams Building on TiDB

Databases for AI Coding Agents

Kimi provisions databases for agent-built applications in about a second, with no idle compute cost.

Persistent Context for AI Agents

Manus uses TiDB Cloud to persist agent context across thousands of stateful iterations per task.

Unstoppable Website Service

Enhancing website resilience with auto-scaling and high availability by moving to the cloud.

The Database

Serverless Auto-Scaling for HTAP Workloads

TiDB Cloud Starter is built for HTAP (Hybrid Transactional and Analytical Processing) combining OLTP and OLAP on the same distributed SQL platform so you can run real-time analytics directly on live operational data.

Built-in columnar storage supports analytical queries on transactional data, without a separate analytics database or ETL pipeline between systems.

Start for Free

Bigger & Better Scale

Fully managed
Scale from 0 to PB without changing your app
Non-blocking schema changes
Support high-performance analytics queries

Faster & Easier Serverless

25 GiB row storage, 25 GiB columnar storage, and 250M Request Units free¹
Only pay for what you use beyond
Scale compute to demand and down to zero when idle
Integrate with modern stack easily like Vercel, ChatGPT, etc.

¹ Monthly total per organization across up to five eligible MySQL-compatible instances. Allowances apply separately to each instance. See pricing details.

Smarter & Scalable

- Leverage vector search to build intelligent AI applications
- Implement Graph-based RAG for high-accuracy results
- Native integrations with leading AI vendors and frameworks
- Keep agent files across sandboxes with TiDB Cloud Filesystem

Full ACID Transactions Without Manual Sharding

TiDB speaks MySQL and PostgreSQL, so you can build with familiar ORMs, and SQL tooling work as-is — most teams migrate without touching their application code. Data is distributed across TiKV nodes automatically. Your application never manages shard keys or routing. Transactions are strongly consistent and ACID-compliant across nodes, and schema changes run online without locking.

Connect with any MySQL-compatible driver or ORM
No shard keys or routing logic in your application
Distributed ACID transactions, strongly consistent
Online DDL: schema changes without downtime or table locks

Branching, Tooling, Monitoring

For Teams Building Agentic Applications

We take pride in our open-source roots and developer community. TiDB Cloud Starter helps you develop queries, test database changes, connect applications to data, and understand database performance as your agentic applications evolve.

40k

GitHub stars

47K

Pull requests

900+

Contributors

AI Assisted SQL Editor

Efficiently and intelligently construct SQL queries. Seamlessly generate intuitive queries, conduct thorough analyses of existing code, and receive tailored function recommendations—all powered by AI.

How to Use AI Assisted SQL Editor

Branching

Manage your database as meticulously as your source code. Effortlessly branch out with CLI, and sync changes with the team effortlessly. Confidently conduct ‘what-if’ experiments in isolated environments ensuring no disruption to the main data stream.

How to Use Branching

Data Service

Experience an all-in-one, low-code solution that streamlines backend development and data access. Our platform offers seamless integration of RESTful APIs and SQL-to-HTTP, simplifying complex processes into efficient, user-friendly operations.

How to Use Data Service

Built-in Monitoring

What keeps you in control. Instantly identify slow queries, resource bottlenecks, and potential issues. Delve deep into SQL statements, visualize emerging trends, and enhance performance using real-time analytics.

How to Use Built-in Monitoring

How to Use AI Assisted SQL Editor

How to Use Branching

How to Use Data Service

How to Use Built-in Monitoring

Vector Search for RAG & AI Apps

Most teams building AI apps end up running a separate vector store alongside their main database. TiDB removes that split. Store embeddings and relational records in the same cluster, query them together with ANN search and SQL joins, and filter by structured metadata in a single statement.

Native vector indexing and ANN search in TiDB
Query vectors and relational data together in standard SQL
No separate vector store to deploy or keep in sync
Works as the retrieval layer for RAG pipelines out of the box

Vector search with TiDB
What is RAG?

Integration

Integrate with Kubernetes & Your Cloud Stack

TiDB Cloud Starter fits naturally into modern, cloud-native environments. It integrates cleanly with:

Kubernetes-based deployment models
CI/CD pipelines
Observability and monitoring stacks
Popular cloud platforms and developer tools

This makes it easy to adopt TiDB as part of an existing infrastructure without disrupting workflows.

See all integrations

Customer Proof: We Build with TiDB

Tens of thousands of developers and startups are building applications with TiDB

Thanks to TiDB, the team debugged with us through the night, fixing issues in minutes. That is the difference between disaster and success.

@Ziming Miao
VP Engineering, Manus

TiDB's architecture, sort of a SQL layer on top of a key value store, was more scalable and more likely to cause fewer problems down the line. With TiDB, we don't need to worry about cross-charge transactions. That's huge.

@Henry Qin
Software Engineer

The ease and comfort of getting started are paramount for us, and the fully managed operations of TiDB Cloud Starter allow us to focus more on our core business, ensuring that our development resources are utilized where they matter most.

@Sky Dong
Founding Engineer, Chaintool

We heavily utilized TiDB's serverless feature to power our backend. This allowed us to scale our backend to handle thousands of requests per second effortlessly while keeping costs low, as we only pay for what we use.

@Eliotte
Hackathon AI Track Winner, Quizmefy

The NO.1 benefit is simplicity. With TiDB, our applications can retain their SQL data model and the ACID guarantees. We don't have to implement any kind of sharding logic, and the database management becomes simpler too.

@Kaustav Chakravorty
Senior Architect, Flipkart

TiDB Cloud Starter is especially beneficial for experimental or early-stage features. It's cost-effective for startups, scalable, and development-friendly.

@Thomas Yu
Founder, KNN3 Networks

With TiDB Cloud Starter, setting up a database is as easy as clicking a button. It handles analytics seamlessly, without the need for manual scaling.

@Godwin
Project Owner of AI-Mon

Scale-in and scale-out in a cluster configuration is also easy, and the ecosystem is very rich, with operation monitoring tools such as Grafana being able to be installed immediately with the installation command (TiUP).

@Kentaro Kitagawa
Senior DBA, LINE Corporation

TiDB Cloud Starter's automatic scaling capabilities allow us to swiftly scale up to meet performance demands and scale down during quieter periods for optimizing costs.

@Yaohui Sun
Data Platform Lead, Chainbase

Start For Free. Pay As You Grow.