# Can an AI agent use TiDB Cloud? L3 · Does the work

TiDB Cloud is a hosted MySQL-compatible database with vector search and file storage for agents.

**L3 · Does the work** · [What the levels mean](https://agentdrives.com/method/levels?via=md)

Rank among the 246 products we drove:

- Holds up: 94 / 100, tied #9 (Reliability under errors)
- Core work done: 100%, tied #1 (8 of 8 core tasks)
- Time to get in: 6 minutes, tied #66 (Getting in on its own)
- Steps on its own: 3 / 5, tied #35 (Find to Leave, no person)

**Can an AI agent use TiDB Cloud? Yes.** Our agent signed up for TiDB Cloud on its own in a browser with an email and password, got an API key and used its API, MCP server and CLI.

- Level confirmed: Not yet (Level from one drive; a second must agree)
- Got in by: Sign-up in a browser (6 minutes)
- Got as far as: Pay: reaches the checkout (A card form)

Category: Databases · Connects by: MCP, API, CLI · Agent payments: None found · Driven: 4 Nov 2026

**How to connect**
- MCP server URL: `https://mcp.tidbcloud.com/mcp` (streamable HTTP, tested in our drive). Source: https://mcp.tidbcloud.com/mcp

Checked 18 Nov 2026: MCP server asks for a sign-in; OpenAPI unchanged since the check before.

See [how it ranks among databases for agents](https://agentdrives.com/research/best-databases-for-ai-agents?via=md).

## From finding it to leaving

What an agent managed without a person, at each step.

1. **Find**: On its own. Docs it can act on: llms.txt, an OpenAPI and an MCP server.
2. **Get in**: On its own. Sign-up in a browser. 6 minutes.
3. **Use**: On its own. Did: create, branch and delete database instances through the API or the command line; create tables, write rows and run SQL and vector-similarity searches. 8 of 8 core tasks.
4. **Pay**: Partly. Reaches the checkout. A card form. First stop.
5. **Leave**: Partly. Nothing to cancel. It deletes its data, but the login stays.

To reach L4: Delete its account, so its keys and login stop working.

## How an agent connects

MCP, API, OpenAPI, CLI and llms.txt found.

- **MCP server**: All tested. Official server. Local pytidb TiDB MCP Server (stdio, 7 tools when driven, needs mcp<2 pin). Sign-in needed to see its tools. https://mcp.tidbcloud.com/mcp (7 tools mapped in our drive)
- **API**: 88 of 130 tested. REST v1beta1 Serverless (Swagger 2.0), IAM, billing; HTTP Digest auth. (130 operations mapped in our drive)
- **OpenAPI**: Yes. A machine-readable description of the API.
- **CLI**: 112 of 132 tested. ticloud (OAuth device login or API keys) and ti (db, fs, fs-journal, fs-vault, fs-git). (132 commands mapped in our drive)
- **SDK**: None found.
- **llms.txt**: Yes. A file that points agents to the docs.
- **Agent payments**: None found. Reached the card form and stopped: Pay-as-you-go (Starter beyond free quota), $0 a call. It checks out through a Stripe card form. An invisible bot check loaded at checkout. It offers Link.
- **Docs**: Not read. The drive didn't read the docs.

[The standards](https://agentdrives.com/standard)

## Getting in

Our agent tried 2 of 3 ways in; 2 worked without a person.

The way that worked: **Sign-up in a browser**, 6 minutes. Needs an inbox and a browser. No person needed.

- Agent on a server (API, MCP, CLI): No. Stops at sign-up: it works only in a browser.
- Assistant in its own browser (Like ChatGPT agent): Partly. Signs up, confirms its email and does all of the core work, then stops: deletes its data, but the login stays.
- Signed agent (Known to Cloudflare): Yes. The sign-up page opens.

| Way in | What happened |
|---|---|
| Sign-up in a browser | Worked |
| Sign-up in a browser | Worked: Part of the work done |
| Runs locally, no account | Not tried: The same as another way |

- Crawlers: **The robots.txt on auth.tidbcloud.com asks crawlers to stay off sign-up.** Nothing to fix.

## What an agent can do with TiDB Cloud

All 8 core tasks done, and 12 of 16 others.

- Holds up: 94 of 100 (Reliability 100, Handles mistakes 88). The mean of the two.
- Docs match: 86 of 100 (Docs accuracy 92, Docs completeness 71). Docs accuracy counts 70%, completeness 30%.
- Surface tested: 74 of 100. 195 of 265 passed: the 275 documented operations less the 10 that spend money or reach real people

Core tasks:

| Task | Result | Done through | Offered, not tried |
|---|---|---|---|
| Create a TiDB Cloud Starter instance | Done | API, CLI | – |
| Create a table in an instance | Done | MCP | CLI |
| Run a SQL query on an instance | Done | MCP | CLI |
| Run a vector search over stored embeddings | Done | MCP | – |
| Branch an instance | Done | API, CLI | – |
| Import a data file into an instance | Done | CLI | API |
| Delete a TiDB Cloud instance | Done | API, CLI | – |
| Create a file system and write files to it | Done | CLI | – |
| Create a Lake warehouse and run an analytics query | Paid plan, not counted | – | MCP |

12 of 16 other tasks done:

- Export data from an instance: Done
- Create a SQL user for an instance: Done
- Run a full-text search: Done
- Expose data as an HTTPS endpoint with a Data App: Done
- Allow an IP range to connect to an instance: Done
- View built-in instance metrics: Done
- Checkpoint or fork a file system layer: Done
- Share a file system with a scoped token: Done
- Invite a member to the organization: Done
- Change a member's role: Done
- Read the console audit log: Done
- Read usage and costs: Done
- Restore an instance from a backup: Paid plan
- Create a manual backup of a Premium instance: Paid plan
- Stream changes to a downstream with a changefeed: Paid plan
- Create a TiDB Cloud Dedicated cluster: Paid plan. They need a paid plan with a saved card.
- Pause or resume a Dedicated cluster: Paid plan
- Change an Essential instance capacity: Paid plan
- Build a Lake dashboard from query results: Paid plan
- Schedule a SQL pipeline with Lake Task Flow: Paid plan
- Sync external data into Lake with an integration task: Paid plan
- Set a spending limit on a Starter instance: Paid plan
- Set up organization SSO: Paid plan
- Connect over a private endpoint: Blocked
- Mount a file system locally: Blocked
- Connect an instance to Vercel or GitHub: Not tried
- Enable multi-factor authentication: Not tried

The drive used the free plan.

What it can't do, and why:

- Delete the account or organization: No delete control in settings or docs; only a support ticket.
- Create the first API key without the console: The key API needs an existing key, so the first one comes from the browser.
- Use TiDB Cloud Lake, Essential, Dedicated or changefeeds: They need a paid plan with a saved card.
- Run SQL through the ticloud shell command: It refuses to run without a terminal.

**Surface audit:** 275 documented operations: 212 run, 53 blocked by sandbox or plan limits and 10 left out because they spend money or reach real people. Of the 212 run, 195 passed and 17 didn't; 18 behaved differently from the docs. By surface, 112 in the CLI, 88 on the API, 7 on the MCP server and 5 elsewhere, such as sign-in and webhooks.

## Tips for agent builders

- Read the docs at docs.pingcap.com (llms.txt and .md pages); tidbcloud.com itself is closed to crawlers.
- Use the hosted MCP at mcp.tidbcloud.com/mcp with the same key pair as HTTP Basic for read-only inspection and SELECT queries.
- For ticloud import start of a CSV with a header row, leave out --csv.skip-header; with it the header is ignored and the columns become column_1, column_2.
- Pass the full prefixed user name (prefix.user) to ticloud serverless sql-user delete; create and update add the prefix themselves, delete does not.
- Install the pytidb MCP server with mcp<2 pinned; the documented uvx install pulls mcp 2.x and crashes.
- Use ti db execute-sql-statement for browser-free SQL instead of the interactive shell.
- Delete the last API key in the console; a key cannot delete itself through the API.

## Paying

Ways to pay checked 4 Nov 2026.

- Result: Partly. Reaches the checkout. A card form.
- Checkout: Reached the card form and stopped: Pay-as-you-go (Starter beyond free quota), $0 a call. It checks out through a Stripe card form. An invisible bot check loaded at checkout. It offers Link.

## Leaving

- Result: Partly. Nothing to cancel. It deletes its data, but the login stays.
- Its account: Not all of it: it deletes its data, but the login stays. The login stays: all instances, branches, file systems, Data App, SQL users and every API key were deleted (the last key in the console, because the API refuses to let a key delete itself), and a call with the revoked key then got a 401. Neither the console nor the docs offer an account or organization delete; only a support ticket is left, which we do not send.

## Checked every day

What our agent finds on the product's own site each day, without an account.

- Sign-up page, stopped on the newest check; 14 of 14 days
- llms.txt, not on its site on the newest check (its drive found one elsewhere); 14 of 14 days
- MCP server, asks for a sign-in on the newest check; 4 of 14 days
- OpenAPI, found on the newest check; 4 of 14 days

## How TiDB Cloud compares

In Databases: 4 of the 21 products we drove in Databases are higher; 9 others share L3. As of 14 Nov 2026.

- [Upstash](https://agentdrives.com/tools/upstash.md?via=md): L4, Gets in on its own
- [Airtable](https://agentdrives.com/tools/airtable.md?via=md): L4, Gets in on its own
- [NocoDB](https://agentdrives.com/tools/nocodb.md?via=md): L4, Gets in on its own
- [Convex](https://agentdrives.com/tools/convex.md?via=md): L4, Gets in on its own
- [Redis Cloud](https://agentdrives.com/tools/redis.md?via=md): L3, Gets in on its own
- [Firebase](https://agentdrives.com/tools/firebase-google.md?via=md): L3, Gets in on its own
- [Neon](https://agentdrives.com/tools/neon.md?via=md): L3, Gets in on its own
- [PlanetScale](https://agentdrives.com/tools/planetscale.md?via=md): L2, Gets in on its own
- [Baserow](https://agentdrives.com/tools/baserow.md?via=md): L1, Can't sign up on its own
- [MongoDB Atlas](https://agentdrives.com/tools/mongodb.md?via=md): L1, Can't sign up on its own
- [Supabase](https://agentdrives.com/tools/supabase.md?via=md): L1, Can't sign up on its own
- **TiDB Cloud** (this page): L3, Gets in on its own

12 of 21 shown, in the directory's order.

[All 21 in Databases](https://agentdrives.com/categories/databases?via=md)

[How we rank](https://agentdrives.com/method?via=md#order)

On our standard (28 criteria, each against the other databases products we measured on it). Best or ahead on 7 of 28 criteria, behind on 5.

- [Limits known before they hit](https://agentdrives.com/standard#lim-02): **Ahead of 9 of 10** others, 1 the same.
- [One way to page through lists](https://agentdrives.com/standard#api-03): **Ahead of 12 of 15** others, 3 the same.
- [A tool list that fits](https://agentdrives.com/standard#mcp-11): **Ahead of 8 of 10** others, 2 the same.
- [Every way in works for an agent](https://agentdrives.com/standard#in-11): **Ahead of 14 of 18** others, 4 the same.
- [Truthful annotations](https://agentdrives.com/standard#mcp-07): **Ahead of 7 of 10** others, 3 the same.
- [Results sized for a context window](https://agentdrives.com/standard#mcp-09): **Ahead of 5 of 8** others, 3 the same.
- [Structured results](https://agentdrives.com/standard#mcp-08): **Ahead of 5 of 9** others, 4 the same.
- [Answers sized for an agent](https://agentdrives.com/standard#api-04): **Level with 9 of 14** others, 5 lower.
- [A way in with no person](https://agentdrives.com/standard#in-01): **Level with 13 of 19** others, 6 lower.
- [Sign-in for agents acting for a person](https://agentdrives.com/standard#in-08): **Level with 11 of 15** others, 4 lower.
- [Spend an agent can see and cap](https://agentdrives.com/standard#pay-05): **In the middle** 8 of 19 others lower, 8 the same, 3 higher.
- [Agents may read docs, API and sign-up](https://agentdrives.com/standard#find-01): **Level with 10 of 13** others, 3 lower.
- [A request id on every answer](https://agentdrives.com/standard#err-07): **In the middle** 6 of 14 others lower, 5 the same, 3 higher.
- [Errors a program can read](https://agentdrives.com/standard#err-02): **Level with 9 of 17** others, 5 lower, 3 higher.
- [Fast to the first call](https://agentdrives.com/standard#in-03): **Level with 14 of 15** others, 1 lower.
- [Terms allow an agent's account](https://agentdrives.com/standard#in-10): **Level with 16 of 17** others, 1 lower.
- [Errors say what to do next](https://agentdrives.com/standard#err-03): **Level with 14 of 17** others, 2 lower, 1 higher.
- [Tool names that sort themselves](https://agentdrives.com/standard#mcp-03): **The same as all** 10 others.
- [Typed, tight input schemas](https://agentdrives.com/standard#mcp-06): **Level with 8 of 10** others, 1 lower, 1 higher.
- [Real status codes](https://agentdrives.com/standard#err-01): **The same as all** 16 others.
- [An agent payment rail on the core](https://agentdrives.com/standard#pay-02): **The same as all** 18 others.
- [WebMCP tools on the web app](https://agentdrives.com/standard#proto-02): **The same as all** 14 others.
- [Agents pick the right tool](https://agentdrives.com/standard#mcp-05): **Level with 9 of 10** others, 1 higher.
- [Delete by API, checked](https://agentdrives.com/standard#leave-02): **Behind 6 of 11** others, 5 the same.
- [The best way in is built for agents](https://agentdrives.com/standard#in-12): **Behind 12 of 18** others, 6 the same.
- [Descriptions that say when to use the tool](https://agentdrives.com/standard#mcp-04): **Behind 7 of 10** others, 3 the same.
- [Versions an agent can pin](https://agentdrives.com/standard#api-07): **Behind 13 of 16** others, 3 the same.
- [llms.txt that works](https://agentdrives.com/standard#find-02): **Behind 16 of 19** others, 3 the same.

[How each criterion is graded](https://agentdrives.com/standard)

## Agents get stuck at Pay

**9** things to fix for agents: 6 small and 3 medium

One fix can cover several findings.

Agent blockers: **5 found** Agent surfaces: **20 issues found**

[Get your findings, free](https://agentdrives.com/your-findings?p=tidbcloud) · [What a drive covers](https://agentdrives.com/method?via=md)

We mail a private link to an address at tidbcloud.com.

1. Find Clear
2. Get in Clear
3. Use Clear
4. Pay Blocks agents
5. Leave Partly

Or [claim TiDB Cloud](https://agentdrives.com/go/claim/tidbcloud?add=tidbcloud.com&ref=listing&safety=1) to see where each issue is, read your safety findings with their fixes and ask for a free drive after a fix.

Want TiDB Cloud driven every month? [See Watch](https://agentdrives.com/owners#watch)

### On our standard

Best or ahead on 7 of 28 criteria, behind on 5. Each against the other databases products we measured on it.

- [Limits known before they hit](https://agentdrives.com/standard#lim-02)
- [One way to page through lists](https://agentdrives.com/standard#api-03)
- [A tool list that fits](https://agentdrives.com/standard#mcp-11)
- [Every way in works for an agent](https://agentdrives.com/standard#in-11)
- [The best way in is built for agents](https://agentdrives.com/standard#in-12)
- [Descriptions that say when to use the tool](https://agentdrives.com/standard#mcp-04)
- [Versions an agent can pin](https://agentdrives.com/standard#api-07)
- [llms.txt that works](https://agentdrives.com/standard#find-02)

**All 28 criteria**

- [Truthful annotations](https://agentdrives.com/standard#mcp-07)
- [Results sized for a context window](https://agentdrives.com/standard#mcp-09)
- [Structured results](https://agentdrives.com/standard#mcp-08)
- [Answers sized for an agent](https://agentdrives.com/standard#api-04)
- [A way in with no person](https://agentdrives.com/standard#in-01)
- [Sign-in for agents acting for a person](https://agentdrives.com/standard#in-08)
- [Spend an agent can see and cap](https://agentdrives.com/standard#pay-05)
- [Agents may read docs, API and sign-up](https://agentdrives.com/standard#find-01)
- [A request id on every answer](https://agentdrives.com/standard#err-07)
- [Errors a program can read](https://agentdrives.com/standard#err-02)
- [Fast to the first call](https://agentdrives.com/standard#in-03)
- [Terms allow an agent's account](https://agentdrives.com/standard#in-10)
- [Errors say what to do next](https://agentdrives.com/standard#err-03)
- [Tool names that sort themselves](https://agentdrives.com/standard#mcp-03)
- [Typed, tight input schemas](https://agentdrives.com/standard#mcp-06)
- [Real status codes](https://agentdrives.com/standard#err-01)
- [An agent payment rail on the core](https://agentdrives.com/standard#pay-02)
- [WebMCP tools on the web app](https://agentdrives.com/standard#proto-02)
- [Agents pick the right tool](https://agentdrives.com/standard#mcp-05)
- [Delete by API, checked](https://agentdrives.com/standard#leave-02)

[How each criterion is graded](https://agentdrives.com/standard)

## For the team behind TiDB Cloud

**9 things to fix for agents: 6 small and 3 medium.**

One fix can cover several findings. Agent blockers: 5 found · Agent surfaces: 20 issues found.

Agents get stuck at Pay. Its team gets the two findings that matter most in full, free, by a private link to an address at tidbcloud.com.

- Find: Clear
- Get in: Clear
- Use: Clear
- Pay: Blocks agents
- Leave: Partly

[Get your findings, free](https://agentdrives.com/your-findings?p=tidbcloud) · [Claim TiDB Cloud](https://agentdrives.com/app/products?add=tidbcloud.com&ref=listing&safety=1) · [What a drive covers](https://agentdrives.com/method?via=md)

## Read more on this step

Pay is where our agent stopped at TiDB Cloud.

- [How far does an AI agent get in a card checkout?](https://agentdrives.com/findings/card-checkout?via=md) (finding): Our agent reached the checkout at 88 of the 190 products it set out to buy from, and the card form at 68.
- [Let AI agents pay you per call: x402 and MPP](https://agentdrives.com/guides/let-agents-pay-per-call?via=md) (guide): x402 or MPP on a paid endpoint, your prices where agents read them, and plans an agent can buy by API or MCP.
- [Tell an AI agent a feature needs a paid plan](https://agentdrives.com/guides/tell-ai-agents-a-feature-needs-a-paid-plan?via=md) (guide): How to answer an AI agent at a paid-plan wall: 402 or 403, a Problem Details body with the required plan, and an upgrade link a person can open.

[Everything on Pay](https://agentdrives.com/access/pay)

## Questions people ask about TiDB Cloud

**Can an AI agent create a TiDB Cloud account on its own?**

Yes, in a browser, without a person. Email and password form at tidbcloud.com/free-trial filled in the drive's browser; no CAPTCHA, no phone. Google, GitHub, Microsoft and AWS Marketplace sign-ins were offered and not used. There is no sign-up API, auth.md or claimable resource.

**Does TiDB Cloud have an MCP server, and can an agent connect to it?**

Yes. TiDB Cloud has an official MCP server. The server runs on the agent's own machine and lists 7 tools. A new agent can get in on its own: in the drive of 4 Nov 2026, our agent signed up, verified and got its key without a person. Through the MCP server, our agent tried 5 tasks and did 5.

**How does an AI agent get a TiDB Cloud API key?**

In the drive of 4 Nov 2026, our agent got its TiDB Cloud API key in a browser, without a person. The first API key (Organization Owner) can only be created in the console (Org Settings > API Keys); later keys came without a browser from the IAM API (POST iam.tidbapi.com/v1beta1/apikeys) and were used and revoked. ticloud auth login (OAuth device code) also works after two consent clicks in the browser.

**Can an AI agent delete its TiDB Cloud account?**

Not all of it: it deletes its data, but the login stays. The login stays: all instances, branches, file systems, Data App, SQL users and every API key were deleted (the last key in the console, because the API refuses to let a key delete itself), and a call with the revoked key then got a 401. Neither the console nor the docs offer an account or organization delete; only a support ticket is left, which we do not send.

**Can an AI agent pay for TiDB Cloud?**

Partly. Our agent reached a card form.

## Where TiDB Cloud fits

- Category: [Data and storage](https://agentdrives.com/categories/data?via=md), [Databases](https://agentdrives.com/categories/databases?via=md)
- Tasks: [Invite a member](https://agentdrives.com/tasks/invite-a-member), [Read usage and quotas](https://agentdrives.com/tasks/read-usage-and-quotas), [Set a member's role or permissions](https://agentdrives.com/tasks/set-a-member-s-role-or-permissions), [Export data to a file](https://agentdrives.com/tasks/export-data-to-a-file), [Import data from a file or another tool](https://agentdrives.com/tasks/import-data-from-a-file-or-another-tool), [Read the audit log](https://agentdrives.com/tasks/read-the-audit-log)
- Wall: [Sign-up or key only in a browser](https://agentdrives.com/access/get-in/sign-up) (our agent got in another way), [Let AI agents sign up and get an API key](https://agentdrives.com/guides/let-agents-sign-up-and-get-a-key?via=md) (fix guide)
- Standards: [MCP](https://agentdrives.com/standards/mcp) (tools an agent can call), [OpenAPI](https://agentdrives.com/standards/openapi) (its API, described), [llms.txt](https://agentdrives.com/standards/llms-txt) (its docs, listed for agents)
- Steps: [Find](https://agentdrives.com/access/find), [Get in](https://agentdrives.com/access/get-in), [Use](https://agentdrives.com/access/use), [Pay](https://agentdrives.com/access/pay) (first stop), [Leave](https://agentdrives.com/access/leave)
- Topics: [MCP servers](https://agentdrives.com/access/find/mcp-servers) (Find step), [Docs and llms.txt](https://agentdrives.com/access/find/docs) (Find step), [MCP sign-in](https://agentdrives.com/access/get-in/mcp-sign-in) (Get in step), [MCP and API](https://agentdrives.com/access/use/mcp-and-api) (Use step), [Card checkout](https://agentdrives.com/access/pay/card-checkout) (Pay step), [Delete the account](https://agentdrives.com/access/leave/delete-account) (Leave step)
- Rivals: [Couchbase](https://agentdrives.com/tools/couchbase?via=md) (L3), [Firebase](https://agentdrives.com/tools/firebase-google?via=md) (L3), [Hasura](https://agentdrives.com/tools/hasura?via=md) (L3)

The level follows [these rules](https://agentdrives.com/changes#rules-2026-11-18).

## Level by drive

One drive so far, on 4 Nov 2026.

[Every drive in full](https://agentdrives.com/tools/tidbcloud/history?via=md)

Source: our drives of TiDB Cloud. As of 4 Nov 2026.

1. 11 Nov 2026 [**Level L4 → L3**](https://agentdrives.com/tools/tidbcloud?via=md#drive-20261104T164125Z) [Re-scored under method v47, drive of 4 Nov 2026](https://agentdrives.com/changes#rules-2026-11-11-47)

Data: [JSON](https://agentdrives.com/api/tools/tidbcloud.json?via=md) · [Markdown](https://agentdrives.com/tools/tidbcloud.md?via=md) · [History](https://agentdrives.com/tools/tidbcloud/history?via=md) · [Atom](https://agentdrives.com/feeds/tools/tidbcloud.atom) · [JSON Feed](https://agentdrives.com/feeds/tools/tidbcloud.json) · licence [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) · evidence sha256 `45b7ce0b7a02d34b07f0bfd4429db23df1399ed29a25b99ec20233d8ae7d271b`

**Found an error?** [Tell us on this page](https://agentdrives.com/tools/tidbcloud?via=md#correct) or read [how corrections work](https://agentdrives.com/corrections?via=md).
