Apper

Apper is a coding tool that builds full-stack apps through the Claude or ChatGPT plan you already pay for. Apper is a vibe-coding app builder that plugs into Claude or ChatGPT, so builds run on the user's existing subscription with no credit meter.

Apper is an AI app builder that avoids a separate credit meter by running on the Claude or ChatGPT subscription a user already holds. Instead of bundling its own model access, it connects to those assistants as a tool (the homepage shows an MCP server URL and a connector listing 28 tools), so the build conversation happens in the user's usual chat and draws on the best model that plan offers. The product targets people who want to describe an app in plain language and end up with something live: founders, agencies, and operators building client portals, booking systems, or billing-enabled tools. The workflow has three stages. The user connects Apper once to ChatGPT or Claude, describes the app in a prompt, then publishes it. In the homepage example, a single request produces database tables with row-level isolation, email and Google sign-in scoped to an organization, Stripe subscription billing with usage caps, and an interface, followed by a preview and a one-click publish to a custom domain. Later prompts can add features such as rescheduling or confirmation emails, and each change can be reviewed before an update goes out. Apper's main argument is about what sits beneath the interface. Many builders generate a polished UI and leave logins, permissions, customer data separation, and payments to whatever the model improvises. Apper instead supplies a tested backend (Postgres, auth, billing, hosting) that exists before the first prompt. It also layers readable permission profiles over database rules, controlling each table, field, and operation, so a faulty AI-written rule cannot open data a profile keeps closed. Each app also ships with connectors and triggers for tools such as Slack, Gmail, Google Sheets, HubSpot, Shopify, Airtable, Trello, and Pipedrive, plus its own API with real keys. That makes it usable with automation platforms like Zapier, Make, and n8n. Among alternatives, it fits users who value backend structure and predictable model costs over a purely UI-first generator, though it depends on holding a paid Claude or ChatGPT plan.