AudioMate — AI Audio Editing SaaS for Creatorshttps://audio-tools.vercel.app

The Problem

Podcasters, radio producers, and content creators spend hours on tedious post-production — cutting filler words, removing silence, reducing noise, mixing in intros and music beds. Most professional tools are expensive, desktop-bound, and have a steep learning curve. AudioMate needed a browser-based platform that automates the boring parts, charges fairly based on actual usage, and stays simple enough for non-engineers to use.

The Approach

I built a full SaaS where every feature lives behind authentication and a metered subscription. Users sign up via Firebase (Google, Facebook, GitHub, or email with verification), pick a plan, and pay through Paddle as merchant of record. Usage is tracked in minutes of processed audio and deducted per job, so billing maps directly to value delivered. Audio editing happens in the browser where possible — waveform trimming, mixing, and MP3 export run client-side — while heavier AI work (transcription for text-based editing, noise reduction) is handled through dedicated API integrations. A scheduled job expires lapsed subscriptions automatically, and an admin area gives oversight of users and plans.

Architecture

The app uses the Next.js App Router with route groups to cleanly separate concerns: public marketing pages, an auth group, and protected private routes for the tools and dashboard. Firebase handles authentication and email verification; signed sessions are issued as JWTs (jose) and verified on protected routes and API calls. MongoDB (via Mongoose) stores users and subscriptions, with a Subscription model tracking plan, allotted minutes, used minutes, and validity window. Payments run through Paddle — transactions are verified server-side against the Paddle API before a subscription is activated or topped up. Each audio job decrements the user's remaining minutes atomically. In the browser, WaveSurfer.js powers waveform visualisation and editing, Tone.js and the Web Audio API handle processing, and lamejs encodes the result to MP3. Speech-to-text (for editing audio by editing its transcript) is powered by AssemblyAI, and noise reduction by Cleanvoice. A cron route sweeps active subscriptions and marks expired ones inactive.

Outcome

Delivered a complete, production-deployed audio SaaS — from sign-up and payments to nine working tools and an admin panel. Users go from raw recording to exported, studio-quality audio entirely in the browser, with fair usage-based pricing across three tiers (free, Mate, Mob). The minutes-metered model ties revenue directly to consumption, and the automated subscription lifecycle keeps billing accurate with no manual intervention.

Tech Stack