velobase/velobase-harness

What it solves

Turning an AI prototype into a commercial product usually requires building repetitive infrastructure from scratch, such as user accounts, payment systems, and usage tracking. Velobase Harness provides a pre-built, open-source foundation that handles these "boring" SaaS requirements, allowing developers to focus exclusively on their unique AI workflows.

How it works

It is an application framework and boilerplate built with Next.js, TypeScript, and React. It integrates a suite of production-ready modules including a credit ledger for usage-based billing, background workers via BullMQ for long-running AI tasks, and anti-abuse tools to prevent free-credit exploitation. The architecture uses a domain-service pattern with a PostgreSQL database and Redis for queuing and rate limiting, and it can be deployed as a single process or split into separate web, worker, and API services.

Who it’s for

• Indie developers who have an AI demo and want to add monetization and production infrastructure quickly.
• Product teams needing to implement usage-based billing, affiliate tracking, and anti-abuse controls for existing AI apps.
• AI-native builders using coding agents (like Cursor or Claude Code) who need a documented production foundation to build upon.

Highlights

• Monetization Suite: Built-in support for subscriptions, usage-based payments, and a credit-based ledger system.
• Growth Tools: Server-side attribution, affiliate/referral programs with double-entry ledgers, and analytics integration.
• Anti-Abuse Controls: Rate limiting, disposable-email checks, and guest quotas to protect expensive AI API costs.
• Production Infrastructure: Includes background workers, multi-LLM chat modules, and lifecycle email systems.