arrow_back Back to Projects
AI SAAS VIDEO PROCESSING

Upload video. Ship it captioned.

Klipify AI needed more than a transcription demo — a complete SaaS for speech transcription, caption timing, styling, and export. Video creators spent hours manually transcribing speech, timing captions, styling subtitles, and exporting files across disconnected desktop tools — per video, per platform, per language. iSoftro engineered an AI captioning SaaS that collapses the whole workflow into one browser tab: upload, auto-transcribe with word-level timing, style on a live preview, and export burned-in video or subtitle files.

person Client

Klipify AI

category Industry

Creator SaaS & Media

verified Services

Product, Next.js + FastAPI

schedule Year

2025

klipify.ai
KLIPIFY Open Studio

Upload video. Ship it captioned.

Word-timed captions
for every clip

Transcribe Now
graphic_eqTranscribe
splitscreenTimeline
paletteStyle
movieExport

Klipify Studio

Transcribe & Style

99.8% Sync Accuracy
SRT+VTT
MP4 Burn-in
description 01

Project Overview

Short-form video runs on captions — most viewers watch muted, and platforms rank captioned content higher. Yet creators serving Nepali, Hindi, and English audiences had no single tool that transcribed accurately, timed precisely, and exported for every platform. Before Klipify AI, captioning meant transcribing by ear in one app, timing in another, styling in a third, and re-encoding with desktop software — with Nepali and Hindi speech barely supported anywhere in the chain. iSoftro stepped in with a full SaaS architecture: a Next.js creator studio, a FastAPI transcription API with a provider abstraction over faster-whisper, background RQ workers for transcription and export, and FFmpeg pipelines that burn pixel-perfect captions into video.

auto_awesome

Key Highlights

  • check_circleSaaS architecture with Next.js + FastAPI and worker fleet
  • check_circleWhisper pipeline tuned for English, Nepali, and Hindi
  • check_circleStudio UX with timeline editing and live preview
  • check_circleFFmpeg export engine for MP4 burn-in + SRT/VTT
crisis_alert 02

The Challenge

The captioning workflow was fractured across four tools and three languages. Every video paid the tax in hours — and multilingual creators paid double. Manual transcription and timing consumed 3+ hours per video, mainstream tools stumbled on Nepali and Hindi speech, hand-nudged timings drifted out of sync, and each platform wanted something different — burned-in MP4 here, SRT there, VTT elsewhere — so creators re-encoded the same captions repeatedly.

lightbulb 03

Our Solution

We designed a creator-first pipeline that maps directly to how short-form videos actually get made: shoot, upload, review words, style, ship. A provider abstraction over faster-whisper transcribes English, Nepali, and Hindi with word-level timestamps, creators edit on an interactive timeline with auto-timing, style with kinetic templates over live preview, and ship via FFmpeg burn-in or SRT, VTT, TXT, and JSON — all on Redis-backed RQ workers that never block the studio.

settings 04

Key Features

graphic_eq
AI Transcription
schedule
Word Timestamps
splitscreen
Timeline Editing
verified
Zero Drift Sync
palette
Kinetic Templates
preview
Live Preview
movie
Burned-in MP4
download
SRT · VTT · JSON
photo_library 05

Website Screenshots

Klipify AI live caption workflow homepage

Upload & Transcribe Queue

Drop a video, pick a language, and watch words appear with timestamps.

Caption Timeline Editor

Caption Timeline Editor

Word-level blocks on a scrubbable timeline with auto-timing that never drifts.

Style Template Studio

Style Template Studio

Kinetic presets and live preview tuned for short-form retention.

Export and Delivery Panel

Export & Delivery Panel

Burned-in MP4 plus every sidecar format, ready for each platform.

Creator Subscription Hub

Creator Subscription Hub

NPR plans, usage quotas, and receipts — billing in the creator’s own currency.

Live Job LedgerLIVE RENDERING
Evening Vlog 02:14 — 214 captions locked. Ad Spot 00:58 — styling. Podcast 05:02 — transcribing. Jobs render while creators keep editing.
Zero Blocked Studios

Live Job Ledger

Jobs render in the background while creators keep editing.

layers 06

Technologies Used

NX
Next.js
database
MySQL
terminal
FastAPI
graphic_eq
Whisper
movie
FFmpeg
deployed_code
Docker

Next.js 16 + TypeScript studio, FastAPI Python backend, MySQL 8.4 + Alembic, faster-whisper pipeline for EN · NE · HI, FFmpeg export engine, Docker + RQ workers with Redis queues.

trending_up 07

Project Results

1,240Videos captioned with word timing
99.8%Sync accuracy, zero drift
3Languages live: EN · NE · HI
“

Captioning went from an evening of typing to a few minutes of review. Nepali and Hindi finally work, sync holds through every edit, and exports are one click for every platform.

KA

Klipify AI

Product Owner, Creator SaaS

rocket_launch

Need a similar AI captioning platform?

Let's build a custom high-performance solution tailored for your creators with transparent sprints.

Next Case Study

Online Exam System: CBT Platform

Timed computer-based testing with automated grading and learning analytics.