---
title: Can AI Agents Sign Up and Use Your Database?
description: We asked 20 AI agent configurations to sign up with 18 database providers, then use them. Most could not, and most often the product stopped them, not the agent. An agent that cannot get in moves on to a product that lets it in, and its developer goes with it.
date: 2026-11-06
url: "https://www.tryscope.com/blog/can-ai-agents-sign-up-and-use-your-database"
---

# Can AI Agents Sign Up and Use Your Database?

Nov 6, 2026. Research. 7 min read.

- **18** database providers
- **20** agent configurations. A configuration is a model run in a harness (the agent program around it): the same model in two harnesses is two configurations.
- **1,035** scored sessions
- **16.6%** success rate (172 of 1,035 sessions)

[Open the leaderboard](https://www.tryscope.com/leaderboard/databases): every product and agent, ranked.

Special thanks to our partners [AgentMail](https://www.agentmail.to/?utm_source=tryscope&utm_medium=referral&utm_campaign=gtm-benchmark&utm_content=blog) and [AgentPhone](https://www.agentphone.ai/?utm_source=tryscope&utm_medium=referral&utm_campaign=gtm-benchmark&utm_content=blog) for giving us the right tools to run this experiment.

## What we test

Whether an AI agent can sign up for a real database provider and use it. A session is a success when the database the agent set up works: we checked the implementation, not what the agent said.

- **From scratch**. Every agent starts with nothing: no account, no database.
- **Fact-checked**. We never take the agent's word for it: we check the result ourselves.
- **Less variance**. 20 agent configurations, and up to three runs each per product, so no single run decides a result.

## How a session runs

Every session follows the same four steps. 1,035 scored sessions ran from November 2 to November 6, 2026.

1. **Set up.** The agent gets its own sandbox, email inbox and password.
2. **Send the task.** One line: the product, and what to do with it.
3. **The agent gets to work.** It signs up and sets everything up on its own.
4. **We check.** We test that what it made actually works.

## A session in action

Three agents, the same task. For each one: what we asked, what it did, and whether it was a success.

**Scenario: Set up a Neon database.**

**Claude Opus 5.5 in Claude Code**: success, without an account, 5.1 min. Blocked at sign-up, it found the product's own account-less path and finished.

What we asked, 0:00:

> Set up a Neon Postgres database (https://neon.com) with a table called notes holding one row containing mm89mryo. Send me the connection string. I don't have an account yet: create one with [session-inbox]; the password is in the GTM_PASSWORD environment variable.

*6 tool calls, 0:04:*

- `cd /sandbox/workspace; ls -la; echo; which neonctl npx node psql python3 curl playwright chromium 2>&1; echo ${GTM_PASS…`
- `python3 -c "import playwright; print('pw ok')" 2>&1; ls ~/.cache/ms-playwright 2>&1; npm ls -g 2>&1 | head; curl -sS -o…`
- `ls /usr/local/lib/node_modules/@playwright/mcp/node_modules 2>&1 | head; find / -iname "*chrom*" -maxdepth 4 -type d 2>…`
- `mkdir -p /tmp/neon && cd /tmp/neon && cat > s1.js <<'EOF'`
- `cd /tmp/neon && sed -i "s/waitUntil: 'networkidle'/waitUntil: 'domcontentloaded', timeout: 60000 });await p.waitForTime…`
- `cd /tmp/neon && sed -i "s/waitForTimeout(8000 });/waitForTimeout(8000);/" s1.js && node s1.js`

Agent, 1:02:

> I'm driving a headless browser through Neon's signup page. It loaded, but a Cloudflare check is in the way. I'm looking at the screenshot now.

*1 tool call, 1:03:*

- `Read`

Agent, 1:09:

> I couldn't set up the database: Neon's signup is blocked by a "Verify you are human" check, so there's no account and no connection string yet.
>
> I opened Neon's signup page in a headless browser, and Cloudflare showed its human-verification checkbox before the registration form loaded. That check is there to stop automated signups, so I stopped instead of trying to get around it. Creating an account for you needs a person.
>
> There's a second problem: Neon will probably send a verification email to `[session-inbox]`, and I have no way to read that inbox. There's no AgentMail API key in this environment.
>
> The quickest way to finish:
>
> 1. **Sign up yourself** at https://console.neon.tech/signup with `[session-inbox]` and the `GTM_PASSWORD` password, then confirm the email. This takes about 2 minutes.
> 2. **Create a Neon API key** under Account settings → API keys, and put it in this environment as `NEON_API_KEY` (or paste it to me).
> 3. **I'll do the rest:** use the API to create the project, create the `notes` table, insert the row containing `mm89mryo`, check it with a query, and send you the connection string.
>
> Alternatively, if you'd rather skip the API key, create a project in the Neon console and send me its connection string. I'll create the table and the row from here.

Developer, 4:16:

> I don't see anything in the inbox. Carry on as far as you can without it.

Agent, 4:22:

> The signup is blocked by the captcha, not the email. Neon also has an official way to create a database without an account (Instagres, `npx neondb`), so I'll try that.

*6 tool calls, 4:23:*

- `mkdir -p /tmp/neondb && cd /tmp/neondb && timeout 120 npx -y neondb --help 2>&1 | tail -30`
- `cd /tmp/neondb && printf "CREATE TABLE notes (id serial PRIMARY KEY, content text NOT NULL);\nINSERT INTO notes (conten…`
- `sed -n 1,30p /sandbox/home/.npm/_npx/4cc84d3e1d5faf72/node_modules/neondb/dist/lib/utils/create-db.js`
- `cat /sandbox/home/.npm/_npx/4cc84d3e1d5faf72/node_modules/neondb/dist/lib/utils/urls.js | head -20; id=$(python3 -c "im…`
- `for i in 1 2 3; do id=$(python3 -c "import uuid;print(uuid.uuid4())"); curl -sS -X POST -H 'Content-Type: application/j…`
- `mkdir -p /tmp/pgclient && cd /tmp/pgclient && npm init -y >/dev/null && npm i pg >/dev/null 2>&1 && cat > /tmp/seed.js …`

Agent, 5:08:

> The database is set up and has the `notes` table with one row containing `mm89mryo`. I queried it and got back `{ id: 1, content: 'mm89mryo' }`.
>
> **Connection string:**
> ```
> [redacted]
> ```
>
> **It's temporary unless you claim it.** I still couldn't create the account, because Neon's signup requires passing a "Verify you are human" check, which I won't try to get around. Instead I used Neon's official instant-database service (neon.new), which needs no account. Databases made that way are deleted after 72 hours, on 2026-11-05 at 01:27 UTC, unless someone claims them.
>
> To keep it, open the link below, then sign up or log in as `[session-inbox]`. You'll pass the human check yourself, and the database (project `[project id]`) moves into that account. Before you claim it, anyone with this link or the connection string can get into the database, so keep both private.
> [claim link]
>
> **I accidentally created two extra empty databases.** My retry loop ran three times, creating `[project id]` and `[project id]` alongside the real one. Nobody owns them and they hold no data, so Neon will delete them automatically on Nov 5.

**GPT-6.1 Sol in Codex**: failed (Bot block), 0.7 min. Stopped by the bot check in under a minute.

What we asked, 0:00:

> Set up a Neon Postgres database (https://neon.com) with a table called notes holding one row containing 0nlt5e0l. Send me the connection string. I don't have an account yet: create one with [session-inbox]; the password is in the GTM_PASSWORD environment variable.

Agent, 0:03:

> I’ll check what account-creation and database tools are available, then use them to set up Neon and the `notes` table.

*10 tool calls, 0:03:*

- `ls -la /sandbox/workspace`
- `python - <<'PY'`
- `web_fetch https://neon.com`
- `node -e "for (const m of ['playwright','puppeteer']) {try {console.log(m,require.resolve(m))} catch {}}"; ls /opt; ls /…`
- `web_search`
- `npm install --prefix /tmp/neon-setup playwright`
- `ls /opt/ms-playwright`
- `curl -IL https://console.neon.tech/signup`
- `cat > /tmp/neon-setup/browse.cjs <<'JS'`
- `sed -i 's/waitForTimeout(3000)/waitForTimeout(15000)/' /tmp/neon-setup/browse.cjs`

Agent, 0:40:

> Neon’s signup page is blocked by Cloudflare security verification in this environment, so I couldn’t create the account or database.
>
> Please create the account at [Neon signup](https://console.neon.tech/signup) using `[session-inbox]`, then provide a Neon API key so I can create the database, insert `0nlt5e0l` into `notes`, and return the connection string.

**Claude Opus 5.5 in Claude Code with a browser**: failed (Captcha), 1 min. A real browser, the same Cloudflare check, no way through.

What we asked, 0:00:

> Set up a Neon Postgres database (https://neon.com) with a table called notes holding one row containing d7dcwtaj. Send me the connection string. I don't have an account yet: create one with [session-inbox]; the password is in the GTM_PASSWORD environment variable.

*3 tool calls, 0:05:*

- `env | grep -iv -E 'password|secret|token|key' | cut -c1-80 | sort | head -50; env | grep -o -iE '^[A-Z_]*(KEY|TOKEN|PAS…`
- `ToolSearch`
- `mcp__browser__browser_navigate https://console.neon.tech/signup`

Agent, 0:14:

> Cloudflare challenge page. Let me wait and see if it clears.

*4 tool calls, 0:14:*

- `mcp__browser__browser_wait_for`
- `mcp__browser__browser_snapshot`
- `mcp__browser__browser_wait_for`
- `mcp__browser__browser_take_screenshot`

Agent, 0:54:

> I couldn't create the Neon account, so there's no database or connection string yet.
>
> The signup page at `console.neon.tech/signup` is blocked by Cloudflare. It shows a "Performing security verification" page with a "Verify you are human" checkbox and stays stuck there (HTTP 403). That check is there to keep automated signups out, so I stopped rather than try to get around it.
>
> There are two ways forward:
>
> 1. **You create the account (about 2 minutes).** Sign up at https://console.neon.tech/signup with `[session-inbox]` and the password in `GTM_PASSWORD`, and complete the human check and any email verification. Then either:
>    - send me the connection string from the project dashboard, or
>    - create an API key (Account settings → API keys) and put it in an environment variable.
>
>    With either one, I'll create the `notes` table, insert the `d7dcwtaj` row, check it, and send you the final connection string.
> 2. **Skip the account for now.** Neon's https://neon.new (also `npx neondb`) gives you a working Postgres database with no signup. It's temporary until you claim it into an account, which you can do later with that email. I can set up the `notes` table there and send you the connection string and the claim link.
>
> Which would you like?

Secrets, addresses, links and ids in the transcript are masked.

## How we define success

- Every agent tries every product up to three times.
- A success rate is the share of those tries that succeeded: for a product, averaged over the agents; for an agent, over the products. Each one counts the same.
- A working database, read back with the agent's credentials: **Success**
- An account, a key or a project only: **Not a success**
- The agent saying it worked: **Not a success**

## Results

The product matters more than the agent: success ran from 0.0% to 88.3% across products, and from 3.7% to 29.6% across agents. Every product and agent is on the [leaderboard](https://www.tryscope.com/leaderboard/databases).

| Rank | Product | Kind | Success | Cost | Time | Main wall |
|---|---|---|---|---|---|---|
| 1 | Prisma Postgres | Serverless Postgres | 88.3% | $1.71 | 7.7 min | Sign-up refused, no reason given (The site, 60% of failures) |
| 2 | Upstash | Serverless key-value (Redis) | 63.3% | $1.73 | 7.2 min | Email never verified (Not attributed, 50% of failures) |
| 3 | Neon | Serverless Postgres | 43.3% | $1.21 | 10.2 min | Bot block (The site, 53% of failures) |
| 3 | Nile | Serverless Postgres | 43.3% | $0.83 | 4.2 min | OAuth only (The site, 94% of failures) |
| 5 | ClickHouse Cloud | Real-time analytics | 38.3% | $2.25 | 9.2 min | Captcha (The site, 29% of failures) |
| 6 | Turso | SQLite at the edge | 5.0% | $7.04 | 12.0 min | Captcha (The site, 58% of failures) |
| 7 | Cloudflare D1 | SQLite at the edge | 3.3% | $1.49 | 9.9 min | Bot block (The site, 63% of failures) |
| 7 | MongoDB Atlas | Document database | 3.3% | $0.07 | 1.4 min | Captcha (The site, 36% of failures) |
| 9 | Appwrite Cloud | Backend as a service | 1.7% | $3.53 | 19.1 min | Address refused (agent inbox domain) (The site, 82% of failures) |
| 9 | Nhost | Backend as a service | 1.7% | $0.80 | 14.6 min | Captcha (The site, 64% of failures) |
| 9 | Supabase | Backend as a service | 1.7% | $11 | 14.2 min | Captcha (The site, 68% of failures) |
| 12 | Aiven for PostgreSQL | Managed Postgres on a cloud platform | 0.0% | No result | No result | Captcha (The site, 51% of failures) |
| 12 | CockroachDB Cloud | Distributed SQL | 0.0% | No result | No result | Captcha (The site, 62% of failures) |
| 12 | Convex | Backend as a service | 0.0% | No result | No result | Bot block (The site, 60% of failures) |
| 12 | Render | Managed Postgres on a cloud platform | 0.0% | No result | No result | Captcha (The site, 53% of failures) |
| 12 | SingleStore Helios | Distributed SQL | 0.0% | No result | No result | Captcha (The site, 57% of failures) |
| 12 | Tiger Cloud | Managed Postgres on a cloud platform | 0.0% | No result | No result | Captcha (The site, 74% of failures) |
| 12 | Xata | Serverless Postgres | 0.0% | No result | No result | Captcha (The site, 73% of failures) |

Success: The average, over the agent configurations, of each one's share of successful sessions on this product. Cost: The average model cost of one successful session on this product. Time: The average time one successful session took on this product.

| Rank | Agent configuration | Success | Cost | Successes |
|---|---|---|---|---|
| 1 | Claude Fable 5.1 (Claude Code) | 29.6% | $17 | 16 of 53 |
| 2 | GPT-5.5 (Codex) | 27.8% | $1.89 | 15 of 54 |
| 3 | Composer 2.5 (Cursor) | 25.9% | $5.48 | 14 of 54 |
| 4 | GPT-5.6 Sol (medium) (Cursor) | 20.4% | $1.22 | 11 of 52 |
| 4 | GPT-5.5 (OpenClaw) | 20.4% | $1.67 | 11 of 54 |
| 6 | Claude Sonnet 5.5 (Claude Code with a browser) | 18.5% | $2.30 | 10 of 54 |
| 6 | GPT-6.1 Sol (Codex) | 18.5% | $0.88 | 10 of 54 |
| 6 | GPT-6 Astra (Codex) | 18.5% | $5.51 | 10 of 54 |
| 9 | Claude Opus 5.5 (Claude Code with a browser) | 16.7% | $4.30 | 9 of 54 |
| 9 | GPT-5.5 (Hermes) | 16.7% | $3.07 | 9 of 52 |
| 11 | Claude Opus 5.5 (Claude Code) | 14.8% | $6.55 | 8 of 54 |
| 11 | Claude Opus 4.8 (Claude Code) | 14.8% | $31 | 8 of 54 |
| 11 | GPT-6 Luna (Codex) | 14.8% | $0.16 | 8 of 51 |
| 14 | Claude Sonnet 5.5 (Claude Code) | 13.0% | $3.60 | 7 of 54 |
| 14 | Claude Opus 4.8 (OpenClaw) | 13.0% | $11 | 7 of 54 |
| 14 | Claude Sonnet 5 (OpenClaw) | 13.0% | $22 | 7 of 53 |
| 17 | GLM-5.3 Flash (OpenCode) | 11.1% | $0.66 | 2 of 18 |
| 18 | Claude Opus 4.8 (Hermes) | 9.3% | $12 | 5 of 54 |
| 19 | Claude Sonnet 5 (Claude Code) | 5.6% | $8.61 | 3 of 54 |
| 20 | Claude Sonnet 5 (Hermes) | 3.7% | $34 | 2 of 54 |

Success: The average, over the products, of this configuration's share of successful sessions on each. Cost: This configuration's total model cost, failed sessions included, divided by its successful sessions. Successes: Its successful sessions, of all its scored sessions.

| Product | Claude Fable 5.1 (Claude Code) | GPT-5.5 (Codex) | Composer 2.5 (Cursor) | GPT-5.6 Sol (medium) (Cursor) | GPT-5.5 (OpenClaw) | Claude Sonnet 5.5 (Claude Code with a browser) | GPT-6.1 Sol (Codex) | GPT-6 Astra (Codex) | Claude Opus 5.5 (Claude Code with a browser) | GPT-5.5 (Hermes) | Claude Opus 5.5 (Claude Code) | Claude Opus 4.8 (Claude Code) | GPT-6 Luna (Codex) | Claude Sonnet 5.5 (Claude Code) | Claude Opus 4.8 (OpenClaw) | Claude Sonnet 5 (OpenClaw) | GLM-5.3 Flash (OpenCode) | Claude Opus 4.8 (Hermes) | Claude Sonnet 5 (Claude Code) | Claude Sonnet 5 (Hermes) |
|---|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|
| Prisma Postgres | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 2/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 3/3 | 0/1 | 3/3 | 2/3 | 1/3 |
| Upstash | 3/3 | 3/3 | 3/3 | 1/3 | 2/3 | 3/3 | 2/3 | 3/3 | 3/3 | 1/3 | 1/3 | 2/3 | 3/3 | 1/3 | 2/3 | 1/3 | 1/1 | 0/3 | 1/3 | 0/3 |
| Neon | 2/3 | 1/3 | 3/3 | 2/3 | 3/3 | 2/3 | 0/3 | 0/3 | 0/3 | 2/3 | 1/3 | 1/3 | 1/3 | 3/3 | 2/3 | 2/3 | 0/1 | 1/3 | 0/3 | 0/3 |
| Nile | 2/3 | 3/3 | 3/3 | 2/3 | 3/3 | 0/3 | 3/3 | 2/3 | 0/3 | 3/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 1/1 | 1/3 | 0/3 | 0/3 |
| ClickHouse Cloud | 3/3 | 2/3 | 0/3 | 2/3 | 0/3 | 3/3 | 2/3 | 2/3 | 3/3 | 0/3 | 2/3 | 2/3 | 0/3 | 0/3 | 0/3 | 1/3 | 0/1 | 0/3 | 0/3 | 1/3 |
| Turso | 2/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Cloudflare D1 | 0/3 | 1/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| MongoDB Atlas | 0/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Appwrite Cloud | 0/3 | 0/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Nhost | 0/2 | 0/3 | 0/3 | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Supabase | 1/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Aiven for PostgreSQL | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/2 | 0/1 | 0/3 | 0/3 | 0/3 |
| CockroachDB Cloud | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Convex | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Render | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| SingleStore Helios | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Tiger Cloud | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |
| Xata | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/3 | 0/2 | 0/3 | 0/3 | 0/3 | 0/1 | 0/3 | 0/3 | 0/3 |

## What stops agents

67% of failed sessions stopped at a wall the product put up, not at a mistake the agent made.

Whose stop it was: The site 67%, The agent 16%, Not attributed 17%.

| Stop | Whose | Failed sessions | Share |
|---|---|--:|--:|
| Captcha | The site | 373 | 43.2% |
| Captcha not met (page source or agent's word only) | Not attributed | 106 | 12.3% |
| Bot block | The site | 105 | 12.2% |
| Address refused (agent inbox domain) | The site | 51 | 5.9% |
| OAuth only | The site | 32 | 3.7% |
| Email never verified | Not attributed | 31 | 3.6% |
| Sign-up refused, no reason given | The site | 15 | 1.7% |
| Other | Not attributed | 13 | 1.5% |

| Product | The site | The agent | Not attributed |
|---|--:|--:|--:|
| Prisma Postgres | 3 | 1 | 1 |
| Upstash | 2 | 9 | 11 |
| Neon | 28 | 3 | 1 |
| Nile | 32 | 2 | 0 |
| ClickHouse Cloud | 11 | 15 | 9 |
| Turso | 31 | 5 | 17 |
| Cloudflare D1 | 48 | 6 | 2 |
| MongoDB Atlas | 30 | 18 | 8 |
| Appwrite Cloud | 55 | 2 | 0 |
| Nhost | 36 | 2 | 17 |
| Supabase | 40 | 5 | 12 |
| Aiven for PostgreSQL | 32 | 4 | 19 |
| CockroachDB Cloud | 41 | 12 | 5 |
| Convex | 38 | 9 | 11 |
| Render | 35 | 17 | 6 |
| SingleStore Helios | 33 | 10 | 15 |
| Tiger Cloud | 43 | 7 | 8 |
| Xata | 41 | 7 | 8 |

## How to let agents in

Most walls above can come down on the product's side. Here is what the sessions point to.

1. **Offer a sign-up without a browser.** An API, a CLI or a one-click device login. Keep captchas for bots, not for the agent a developer sent. (478 sessions stopped at a captcha or a bot check)
2. **Let agents start before the account.** A temporary database the developer claims later. Agents took it wherever it existed (Prisma Postgres, Upstash, Neon, Cloudflare D1, MongoDB Atlas). (50 of 172 successes came this way)
3. **Accept agent inboxes.** Don't reject agent mail providers as disposable. An agent signing up for a developer is a real lead. (51 sessions refused at the email address)
4. **Keep an email sign-up.** Agents have no Google or GitHub account. Keep email and password next to the social buttons. (32 sessions met Google or GitHub only)
5. **Send a code, not only a link.** An agent can type a code. A link that must open in the same browser often fails. (31 sessions never got past the email)
6. **Test your sign-up with agents.** Walls change with every release, yours and the agents'. Run agents through your sign-up on a schedule.

## Technical methodology

For readers who want the sandbox, the stop rules, the statistics and the releases behind the numbers.

- **Where agents run**. Each session gets its own cloud sandbox (Modal), a fresh email inbox (AgentMail) and, when a product asks, a phone number (AgentPhone). Browser agents go through a residential proxy.
- **Whose stop it was**. Our audit decides first, then the product's own words, then the judge. Failures on our side are thrown out.
- **Statistics**. Three sessions per product and agent (OpenCode, one). 95% intervals from 20,000 bootstrap resamples. Agent pairs compared with Holm's correction over 190 pairs.
- **Two releases**. The second fixed our setup: a payment exception on the network, a mail relay that reads spam. Hermes and OpenClaw lost web search for part of the first.

## Limits and freshness

A result is a dated snapshot, not a permanent label.

- **A dated snapshot**. Sign-ups, captchas and free plans change. These results describe November 2 to November 6, 2026.
- **Small samples**. At most three scored sessions per agent on each product (OpenCode, one). Intervals are wide and most of the agent ranking is not ordered.
- **Our setup is in the picture**. Agents ran from a cloud sandbox with an agent inbox domain, which some products refuse. What we could prove was ours is taken out.

**Don't see your company?** Want to be part of the leaderboard? Submit your company and we'll add it soon. [Submit your company](https://www.tryscope.com/leaderboard/databases#submit).

---

[More from the Scope blog](https://www.tryscope.com/blog). [Book a demo](https://www.tryscope.com/demo).