Airtable vs Convex for AI agents

As of 15 Oct 2026, Airtable gets an AI agent further: L4 · Runs on its own against L3 · Does the work.

AirtableL4Runs on its own

Find: Some docs for agents: llms.txt and an MCP server.
Access: MCP sign-in, sign-up in a browser.
Use: 100% of the work.
Pay: Reaches the checkout: Stopped at the $24 a month charge.
Leave: Leaves on its own: Nothing to cancel, deletes what it made.

Through every checkpoint. Next: pay without a person

Gets in on its own today, L4 · Runs on its own. Sign-up blocks headless requests, but our agent reached the browser form with no visible challenge and confirmed its email there. Personal access tokens, the first workspace and turning on automations all still need that browser UI. Once past those steps, the API and the official MCP server ran almost all of Airtable's core work: bases, records, schema, webhooks, automations, interfaces and forms. Upgrading opens a Stripe card form inside the app; our agent found no agent payment rail and stopped before paying. It started no plan, so there was nothing to cancel, and it deleted its account through account settings.

Gets furtherDriven 15 Oct 2026 · The drive and its evidence

ConvexL3Does the work

Find: Docs for agents: llms.txt, an OpenAPI and an MCP server.
Access: Runs locally: Can't sign up for one.
Use: 70% of the work.
Pay: Payment not reached.
Leave: Leaves on its own: Nothing to cancel, made nothing to delete.

Got in; payment not reached. Next: let an agent sign up for an account on its own (without one it does 70% of the core work)

Gets in on its own, L3 · Does the work: a documented no-account route provisions a local deployment, where our agent pushed schema and functions, inserted and patched documents, ran an HTTP action, a cron and a scheduler, and set env vars, each checked back. The cloud side stopped at sign-up: a visible Cloudflare "Verify you are human" challenge blocked account creation, so no cloud project, deploy key or MCP session followed. Core backend work, plus file storage, text and vector search, custom-JWT auth and live subscriptions, ran on the local deployment; cloud paying and account deletion stayed out of reach, since both sit behind that same human checkpoint and the dashboard.

Driven 2 Oct 2026 · The drive and its evidence

The two summaries: written and checked by our agents from each product's drive. How we write.

Where each one stops.

AirtableThrough every checkpoint

ConvexGot in; payment not reached

Airtable and Convex, checkpoint by checkpoint
Checkpoint	Airtable	Convex
Docs	Headless	Headless
Sign up	In a browser	Not needed
Verify	In a browser	Not needed
Key	In a browser	Not needed
First call	Headless	Headless
First action	Headless	Headless
Pay	In a browser	Not reached
Cancel	Not needed	Not needed
Delete	In a browser	Blocked

Headless or in a browser.

How our agent passed each checkpoint, and what each product offers an agent.

Checkpoints by route, API, MCP, CLI and SDK

Checkpoints by how our agent passed them, and what each product offers an agent
	Airtable	Convex
Headless (checkpoints)	3	3
In a browser (checkpoints)	5	0
Needs a human (checkpoints)	0	0
Blocked (checkpoints)	0	1
Not tested (checkpoints)	0	1
Work done	100% of the core	70% of the core
API	Yes	Yes
OpenAPI	No	Yes
MCP server	Yes (official)	Yes (official)
CLI	Yes	Yes
SDK	–	Yes
llms.txt	Yes	Yes

Bot checks, and where a person was needed.

Walls our agent met
	Airtable	Convex
Bot check	HUMAN let it through at sign up	Cloudflare human check on the sign-up page; it got in another way
Wall type	No bot-check wall	A check of our network and browser
First wall	None	Got in; payment not reached
A person helped	No	No

A check of our network and browser: It judged where our agent connects from and its browser, not a person. An agent that works on its own from a data center doesn't get through unless the owner lets signed agents in. How agents in a person's own browser fare isn't measured yet. Where we drive from.

Level and part scores.

Airtable and Convex: level and part scores, 0 to 100
	Airtable	Convex
Level	L4 · Runs on its own	L3 · Does the work
Holds up	78 of 100, fair	Not measured
Agent-ready setup	75 of 100, fair	100 of 100, good
Standards	MCP spec of 25 Nov 2025 (1 version behind) · MCP discovery (behind) · MCP client registration (current)	None recorded

Docs match is measured in an Audit. What the scores mean.

Pay, cancel and delete.

Paying, cancelling and deleting the account
	Airtable	Convex
Pay	Reaches the checkout	Payment not reached
Free plan or trial	Trial, no card	Free plan
Cancel	Not needed	Not needed
Delete the account	In a browser	Blocked

How each one takes payment

Airtable: Reaches the checkout: stopped at the $24 a month charge.
Convex: Payment not reached.

Airtable vs Convex: the short answers.

Which can an AI agent use, Airtable or Convex?

Airtable gets an AI agent further: L4 · Runs on its own against L3 · Does the work. With Airtable our agent went through every checkpoint; with Convex it got in (payment not reached). Driven 15 Oct 2026 (Airtable) and 2 Oct 2026 (Convex).

Can an AI agent sign up for Airtable and Convex on its own?

Airtable: yes, in a browser. Convex: not needed.

Do Airtable and Convex have an MCP server?

Airtable: yes (official). Convex: yes (official).

Can an AI agent delete its Airtable or Convex account?

Airtable: yes, in a browser. Convex: blocked.