Our own product2026Viralance
An AI video creation studio for web and iOS.
- Project
- Our own product
- Our role
- Product · Full stack · AI infrastructure · Payments
- Duration
- 6 months
01 / PRODUCT & APPROACH
A whole platform behind one create button.
Viralance is our own product for creating AI-assisted video from an idea or image. Web and iOS experiences sit alongside model selection, generation tracking, credits and content management as parts of one product.
In AI video generation the interface is the easy part. Generation takes minutes and does not fit in one HTTP request; more than 16 models want different durations, resolutions and input modes; a failed job needs a credit refund and the same refund must never apply twice. Then there is abuse: fake accounts, disposable emails, content outside policy.
From idea to content
- 01
Define the idea
Start with a prompt, source image or ready-to-use preset.
- 02
Track generation
Choose a supported duration and format, then follow the job in the video library.
- 03
Manage the output
Find finished content in your library, download it and move it into publishing.
DECISIONS THAT SHAPED THE PRODUCT
Waiting is part of the product experience.
Video generation does not finish in a single request. The interface connects to a job tracked by its ID, and users move to their video library after starting it. Models, durations, resolutions and input modes are defined through a shared catalog.
Credits stay consistent when jobs fail.
Generation failure, repeated status checks and deletion of a pending job can affect the same credit record. Shared refund rules and state tracking prevent the same job from being refunded more than once. Payment and generation lifecycles are handled together.
Web and mobile, each in its own context.
Model and content management are prominent on the web. On iOS, discovery, video presets and selecting images from the phone take priority. The same product idea is expressed through workflows suited to each device.
THE RESULTING PRODUCT
Generation, payments and mobile in one product.
Live on web and iOS; credit packs are one-time purchases through Stripe. The same generation backend is exposed as an MCP server and used across the studio’s other projects (game icons, art assets). The web side leads with model and content management, iOS with discovery and ready-made presets.
02 / Inside the product
The experience, on screen.
Open a screen to explore the details.08 images
03 / Scope & details
The details make the experience.
- 01
Job queue with observable state
- 02
Multiple video models through a shared catalog
- 03
State-based credit rules that prevent duplicate refunds
- 04
Bulk generation, scheduling and TikTok publishing
- 05
Video creation and ready-to-use presets in the iOS app
Technology
Next.js / Firebase / Cloud Functions / Stripe / React Native
INSIDE THE PRODUCT
The engineering behind the interface.
Systems we built and concrete problems we addressed in this project.
Managing generation as a job lifecycle
Video generation is tracked through pending, processing and completed jobs. The interface and background generation tasks share a state model, rather than depending on one long HTTP request.
Preventing repeated credit refunds
Deletion, polling and scheduled processing check the same refund state. The decision depends on job status, credit cost and whether a refund already occurred; the logic is isolated in a module with unit tests.
The abuse and moderation layer
The abuse-controls, abuse-policy, moderation-core, payment-guard, disposable-domains, turnstile and watermark modules stop fake accounts, disposable emails and off-policy content on the server; openai-proxy keeps the key in one place. The same backend is exposed as an MCP server through the viralance-mcp package.
Your idea could be next.
What could we build for you?
From a first release to a growing platform. Let’s plan the design, development and launch together.
From Türkiye to the US, UK, Australia and everywhere in between.
ENGINEERING NOTES FROM THIS PROJECT
Viralance: the job lifecycle behind video generation ↗