Claimable Postgres

Claimable Postgres

Instant Postgres databases for local development, demos, prototyping, and test environments. No account required. Databases expire after 72 hours unless claimed to a Neon account.

Quick Start

curl -s -X POST "https://pg.new/api/v1/database" \
  -H "Content-Type: application/json" \
  -d '{"ref": "agent-skills"}'

Parse connection_string and claim_url from the JSON response. Write connection_string to the project's .env as DATABASE_URL.

For other methods (CLI, SDK, Vite plugin), see Which Method? below.

Which Method?

• REST API: Returns structured JSON. No runtime dependency beyond curl. Preferred when the agent needs predictable output and error handling.
• CLI (npx get-db@latest --yes): Provisions and writes .env in one command. Convenient when Node.js is available and the user wants a simple setup.
• SDK (get-db/sdk): Scripts or programmatic provisioning in Node.js.
• Vite plugin (vite-plugin-db): Auto-provisions on vite dev if DATABASE_URL is missing. Use when the user has a Vite project.
• Browser: User cannot run CLI or API. Direct to https://pg.new.

REST API

Base URL: https://pg.new/api/v1

Create a database

curl -s -X POST "https://pg.new/api/v1/database" \
  -H "Content-Type: application/json" \
  -d '{"ref": "agent-skills"}'

The connection_string returned by the API is a pooled connection URL. For a direct (non-pooled) connection (e.g. Prisma migrations), remove -pooler from the hostname. The CLI writes both pooled and direct URLs automatically.

Response:

{
  "id": "019beb39-37fb-709d-87ac-7ad6198b89f7",
  "status": "UNCLAIMED",
  "neon_project_id": "gentle-scene-06438508",
  "connection_string": "postgresql://...",
  "claim_url": "https://pg.new/claim/019beb39-...",
  "expires_at": "2026-01-26T14:19:14.580Z",
  "created_at": "2026-01-23T14:19:14.580Z",
  "updated_at": "2026-01-23T14:19:14.580Z"
}

Check status

curl -s "https://pg.new/api/v1/database/{id}"

Returns the same response shape. Status transitions: UNCLAIMED -> CLAIMING -> CLAIMED. After the database is claimed, connection_string returns null.

Error responses

CLI

npx get-db@latest --yes

Provisions a database and writes the connection string to .env in one step. Always use @latest and --yes (skips interactive prompts that would stall the agent).

Pre-run Check

Check if DATABASE_URL (or the chosen key) already exists in the target .env. The CLI exits without provisioning if it finds the key.

If the key exists, offer the user three options:

1. Remove or comment out the existing line, then rerun.
2. Use --env to write to a different file (e.g. --env .env.local).
3. Use --key to write under a different variable name.

Get confirmation before proceeding.

Options

Alternative package managers: yarn dlx get-db@latest, pnpm dlx get-db@latest, bunx get-db@latest, deno run -A get-db@latest.

Output

The CLI writes to the target .env:

DATABASE_URL=postgresql://...              # pooled (use for application queries)
DATABASE_URL_DIRECT=postgresql://...       # direct (use for migrations, e.g. Prisma)
PUBLIC_POSTGRES_CLAIM_URL=https://pg.new/claim/...

SDK

Use for scripts and programmatic provisioning flows.

import { instantPostgres } from 'get-db';

const { databaseUrl, databaseUrlDirect, claimUrl, claimExpiresAt } = await instantPostgres({
  referrer: 'agent-skills',
  seed: { type: 'sql-script', path: './init.sql' },
});

Returns databaseUrl (pooled), databaseUrlDirect (direct, for migrations), claimUrl, and claimExpiresAt (Date object). The referrer parameter is required.

Vite Plugin

For Vite projects, vite-plugin-db auto-provisions a database on vite dev if DATABASE_URL is missing. Install with npm install -D vite-plugin-db. See the Claimable Postgres docs for configuration.

Agent Workflow

API path

1. Confirm intent: If the request is ambiguous, confirm the user wants a temporary, no-signup database. Skip this if they explicitly asked for a quick or temporary database.
2. Provision: POST to https://pg.new/api/v1/database with {"ref": "agent-skills"}.
3. Parse response: Extract connection_string, claim_url, and expires_at from the JSON response.
4. Write .env: Write DATABASE_URL=<connection_string> to the project's .env (or the user's preferred file and key). Do not overwrite an existing key without confirmation.
5. Seed (if needed): If the user has a seed SQL file, run it against the new database:

   psql "$DATABASE_URL" -f seed.sql
   

6. Report: Tell the user where the connection string was written, which key was used, and share the claim URL. Remind them: the database works now; claim within 72 hours to keep it permanently.
7. Optional: Offer a quick connection test (e.g. SELECT 1).

CLI path

1. Check .env: Check the target .env for an existing DATABASE_URL (or chosen key). If present, do not run. Offer remove, --env, or --key and get confirmation.
2. Confirm intent: If the request is ambiguous, confirm the user wants a temporary, no-signup database. Skip this if they explicitly asked for a quick or temporary database.
3. Gather options: Use defaults unless context suggests otherwise (e.g., user mentions a custom env file, seed SQL, or logical replication).
4. Run: Execute with @latest --yes plus the confirmed options. Always use @latest to avoid stale cached versions. --yes skips interactive prompts that would stall the agent.

   npx get-db@latest --yes --ref agent-skills --env .env.local --seed ./schema.sql
   

5. Verify: Confirm the connection string was written to the intended file.
6. Report: Tell the user where the connection string was written, which key was used, and that a claim URL is in the env file. Remind them: the database works now; claim within 72 hours to keep it permanently.
7. Optional: Offer a quick connection test (e.g. SELECT 1).

Output Checklist

Always report:

• Where the connection string was written (e.g. .env)
• Which variable key was used (DATABASE_URL or custom key)
• The claim URL (from .env or API response)
• That unclaimed databases are temporary (72 hours)

Claiming

Claiming is optional. The database works immediately without it. To optionally claim, the user opens the claim URL in a browser, where they sign in or create a Neon account to claim the database.

• API/SDK: Give the user the claim_url from the create response.
• CLI: npx get-db@latest claim reads the claim URL from .env and opens the browser automatically.

Users cannot claim into Vercel-linked orgs; they must choose another Neon org.

Defaults and Limits

Region cannot be changed for claimable databases. Unclaimed databases have stricter quotas. Claiming resets limits to free plan defaults.

Auto-provisioning

If the agent needs a database to fulfill a task (e.g. "build me a todo app with a real database") and the user has not provided a connection string, provision one via the API and inform the user. Include the claim URL so they can keep it.

Safety and UX Notes

• Do not overwrite existing env vars. Check first, then use --env or --key (CLI) or skip writing (API) to avoid conflicts.
• Ask before running destructive seed SQL (DROP, TRUNCATE, mass DELETE).
• For production workloads, recommend standard Neon provisioning instead of temporary claimable databases.
• If users need long-term persistence, instruct them to open the claim URL right away.
• After writing credentials to an .env file, check that it's covered by .gitignore. If not, warn the user. Do not modify .gitignore without confirmation.

When to Use

Use this skill when tackling tasks related to its primary domain or functionality as described above.

AGI Framework Integration

Adapted for @techwavedev/agi-agent-kit Original source: antigravity-awesome-skills

Memory-First Protocol

Retrieve prior design decisions (color palettes, typography, spacing scales) to maintain visual consistency across sessions. Cache generated design tokens.

# Check for prior frontend/design context before starting
python3 execution/memory_manager.py auto --query "design system decisions and component patterns for Claimable Postgres"

Storing Results

After completing work, store frontend/design decisions for future sessions:

python3 execution/memory_manager.py store \
  --content "Design system: adopted 8px grid, Inter font family, HSL color tokens with dark mode support" \
  --type decision --project <project> \
  --tags claimable-postgres frontend

Multi-Agent Collaboration

Share design decisions with backend agents (API contract changes) and QA agents (visual regression baselines).

python3 execution/cross_agent_context.py store \
  --agent "<your-agent>" \
  --action "Implemented UI components — new design system with accessibility compliance (WCAG 2.1 AA)" \
  --project <project>

Design Memory Persistence

Store design system tokens and component decisions in Qdrant so any agent on any platform (Claude, Gemini, Cursor) can retrieve and apply consistent styling.