Free | Lisa Lipps Golden Lipps Full Sceneupscale

| Goal | Metric | Target (12 mo) | |------|--------|----------------| | User acquisition – attract new users with a zero‑cost entry point | New free‑tier accounts | 150 k | | Conversion – move free users to paid plans | Free‑to‑paid conversion rate | 8 % | | Retention – keep users coming back for repeat upscales | Monthly active free users (MAU) | 45 % of sign‑ups | | Brand awareness – position Golden Lipps as the go‑to “scene‑wide” upscale tool | Social mentions / press hits | 30 % YoY increase | | Technical reliability – deliver upscale results in < 10 s for typical 1080p input | 99 % of jobs completed within SLA | 99 % |

| Method | Endpoint | Auth | Request | Response | |--------|----------|------|---------|----------| | POST | /v1/upscale/fullscene | API‑Key (free) | multipart/form-data
• file (image/video)
• scale (2, 3, 4)
• preserve_color (bool) | 202 Accepted
"job_id": "uuid", "status_url": "/v1/jobs/id" | | GET | /v1/jobs/job_id | API‑Key | – | processing |

Rate limits: 5 images / day / API‑Key; 1 video (≤ 30 s) / day. Exceeding returns 429 Too Many Requests. lisa lipps golden lipps full sceneupscale free

| Service | Tech Stack | Key Responsibilities | |---------|------------|-----------------------| | Web Front‑End | React 18 + Vite, TailwindCSS | UI, client‑side validation, temporary UUID handling | | API Gateway | Node.js (Express) + TypeScript, OpenAPI spec | Public endpoints, rate limiting, request routing | | Auth Service | Auth0 (or custom JWT) | Temporary UUID issuance, optional user accounts | | Job Queue | RabbitMQ (or AWS SQS) | Reliable async job dispatch, back‑pressure handling | | Worker Pool | Python 3.11 + FastAPI, Docker, CUDA 12 | Pull jobs, run upscale model, store result | | Upscale Engine | Custom TensorRT‑optimized model (based on ESRGAN/Real‑ESRGAN) + OpenVINO fallback | Full‑scene processing, 4× upscale, color‑preservation mode | | Object Storage | MinIO (self‑hosted) or AWS S3 | Persist original & upscaled assets, TTL for free results (48 h) | | Metrics & Logging | Prometheus + Grafana, ELK stack | SLA monitoring, usage analytics, error tracing |