Supapool

Supapool is an AI coding infrastructure tool for teams that run multiple coding agents against Supabase-backed applications. It gives each agent a real, isolated Supabase environment instead of asking developers to share one local database, mock cloud services, or wait for slow branch creation. The workflow is intentionally simple: wrap a development or test command with `npx @supapool/cli run --`, and Supapool leases a clean instance, injects Supabase credentials, keeps the lease alive while the command runs, and releases the instance when the process exits. The product is built for the new reality of parallel agent work. When several agents operate in separate worktrees, a shared database becomes a bottleneck. One agent can wipe test data, run migrations, or change auth state while another agent is still using it. Supapool solves that by giving every run its own disposable Postgres, Auth, Realtime, and S3-compatible Storage stack. That makes integration tests and migration checks closer to production than mocks, while keeping production and shared staging data untouched. Supapool is CLI-first. There is no dashboard-centered workflow; accounts, usage, leases, and costs are exposed through the CLI and API so agents and CI jobs can consume them directly. The CLI mirrors standard Supabase variables such as `SUPABASE_URL`, `SUPABASE_ANON_KEY`, `SUPABASE_SERVICE_ROLE_KEY`, and `DATABASE_URL`, plus common framework-specific environment names for Next.js, Vite, Astro, Svelte, Expo, Create React App, Gatsby, Nuxt, Prisma, and Postgres clients. Secret keys are not written into browser-public variables. Each instance is leased with a default 30-minute TTL. The CLI renews it every five minutes while the wrapped command is alive. When the command exits, or if renewals stop, the lease expires and the slot is wiped and returned to the pool. That disposable lifecycle makes Supapool useful for coding agents, pull-request checks, preview workflows, and local development sessions where a real backend is needed briefly. Pricing is currently beta-oriented rather than a published tier table. The official site says Supapool is free during beta and asks users to contact the team by email. For production buyers, that means the safe assumption is beta access now and custom or future usage-based pricing later. Supapool is best for developers already using Supabase who want safer parallel agent runs without running several heavy local Supabase stacks on one laptop. For builders, the main advantage is speed with lower coordination risk. A coding agent can run migrations, seed fixtures, call authenticated endpoints, and exercise storage flows against a real service boundary. Another agent can do the same in a separate lease. Humans can review failures without wondering whether a shared test database was already changed by a different worker. Supapool also fits pull-request automation: a CI job can acquire a short-lived Supabase instance, run tests against the migrated schema, and discard every change at the end. That makes it a focused infrastructure layer for teams moving from single-agent experiments to parallel agent development.