I'm a freelance AI Video & Image Creator, Visual Designer, and Workflow Consultant based in Tripura, India, and I work with clients all over the world.

Think of me as a full production studio in one person. I take a project from the first spark of an idea all the way to the finished result, and I look after every step in between. That might be a cinematic commercial campaign, a fresh brand identity, or a short film made entirely with AI. I also help companies bring AI into their everyday work and get their teams genuinely comfortable using it.

Real art direction, with modern tools doing the heavy lifting behind it.

🎬 Cinematic AI Video & Commercial Campaigns Ad concepts and short films that actually look real, the kind of visuals that make a brand look world class.

🎨 Visual Design & Brand Identity Logos, social grids, and full brand looks that give a business a clear, memorable style.

🤖 AI Training & Workflow Consulting I help teams fold AI into how they already work, then teach them to run with it on their own.

💻 Interactive Web Experiences Turning a creative idea into a website or app that people genuinely enjoy using.

• House of Shafaq (Dubai) · Luxury fashion visuals and creative direction
• Glitchberry (Bangalore) · Visual storytelling and digital aesthetics
• Hornbill Restaurant (Agartala) · Full brand identity and social media design
• Fiverr · Top rated freelancer delivering campaigns for clients worldwide

"The tools change almost every week, so the most useful thing I own is how fast I can learn."

I don't really wait around for the perfect moment or a formal course. I just start, learn fast, and turn a messy brief into a finished piece.

Whether you need an ad campaign that stops the scroll, a fresh brand identity, or someone to help your team get good at AI, I would love to hear from you.

📩 Email: madebysaira@proton.me 🌐 Portfolio: madebysaira.me 📅 Book a quick call: available through my website

I open-source my production workflows and prompt libraries to help fellow creators:

• uiux-design-skills — Premium dark design system as a reusable skill: flat solid colors, glassmorphism cards, Poppins, Font Awesome, real-time metric patterns, plus keyless curl recipes for icons, images, video, and fonts.
• PromptLab — Browser-based lab for testing AI video prompts one variable at a time. Lock blocks, get clean-test verdicts, and copy model-formatted prompts for Kling, Veo, Runway, Luma, and Seedance. No install, no API keys, works offline.
• AIVideoQualityGate — Offline ffprobe and ffmpeg checks for AI video renders before client delivery. Catches missing audio, wrong formats, frozen sections, and loudness problems.
• AIVideoAdherenceGate — Post-render semantic + motion health gate. Checks whether a rendered clip actually honored the creative contract and is physically healthy (motion, morph-drift, lip-sync). Completes the AIVideo* safety net with CreditGuard + QualityGate. Offline-first, optional vision scoring. Also shipped as an xyOps plugin (v1.0.1 — audited + fixed). Visual guide: adherence-gate.madebysaira.me.
• AIVideoRunbook — The conductor for the safety net: a lightweight, offline-first CLI that runs CreditGuard → QualityGate → AdherenceGate in order, logs every clip to a local SQLite catalog (provider, model, seed, cost, verdict), and emits exit codes 0/1/2 for n8n/xyOps IF-nodes. Find any render later by prompt, provider, or verdict. The logbook that turns three gates into one credit-to-delivery pipeline.
• AIVideoDiagrams — A 14-diagram design system for AI video client work (pipeline to funnel), plus a matching pipeline diagram for AIVideoRunbook. One design system for both. Plain SVG, light and dark, no runtime.
• MotionGraphicsStudio — Complete Claude Co-work setup for AI-powered motion graphics. Style sheets (Vox, Ink Vector), reusable skills, and production templates. Clone and start creating in 10 minutes.
• PostProductionForAIVideo — Post-production templates and scripts for AI video polish (Remotion, ffmpeg, AE). Client-ready workflows that save hours per video.
• VideoWorkflows — Practical n8n pipelines and prompt kits for turning ideas into finished AI video without the manual grind. Full automation from schedule to Drive + Sheets.
• RemotionPlaybook — Recreate any brand video with AI agents (Claude Code + Gemini + Remotion). Scene-split scripts, agent prompts, and the grading rubric, all runnable.
• cinematic-ai-prompts — Curated prompt library for commercial AI video (Kling, Veo, Runway, Luma)
• n8n-ai-creator-pipelines — Production-ready n8n workflows for AI studios (brief intake, batch generation, client review, delivery)
• ai-motion-graphics-comparison — Model benchmarks for AI motion graphics (Omni Flash vs Remotion vs others). Decision tree, prompt templates, hybrid workflows.
• CharacterConsistency — Practical toolkit for locking characters across shots in commercial AI video. Locked style blocks, decision tree, model-specific prompts, full case studies.
• CharacterBibleGenerator — A model-agnostic character bible, reference sheet workflow, and frame review checklist for recurring AI video characters.
• AIVideoCreditGuard — Offline pre-generation guard that lints prompts, estimates real cost (sticker × retries), and flags morph/consistency risks before Kling, Veo or Runway burns a credit. Companion to AIVideoQualityGate.
• AIVideoBatchQueue — Crash-safe, resumable batch queue for AI video generation jobs. One SQLite file of truth: atomic CAS claims (no double-runs), attempt-tracked retries, a hard credit budget ceiling (jobs that would overspend are skipped, never billed), and export --jsonl provenance. Provider-agnostic command templates — Kling, Veo, Runway, ComfyUI, or a shell one-liner. Zero required dependencies, Python 3.10+. 5th pillar of the AIVideo safety net.
• ComfyUI-VideoGate — Offline post-render video QC gate that runs inside ComfyUI graphs. Four nodes (Probe/Check/Report/Enforce) turn "the render finished" into "the render is deliverable": ffprobe facts, delivery profiles (client_delivery, social_vertical, archival_master), black-frame / frozen-frame / loudness checks, PASS/WARN/FAIL verdicts, reports, and an Enforce node that hard-stops the workflow on broken clips. No GPU, no cloud. 6th pillar of the AIVideo safety net.

New repos weekly! Star them if you find them useful!

Writing about the work as I ship it:

• The Gate That Was Missing: Did Your Clip Actually Do What You Asked? — August 2026
• Before You Burn the Credit: A Pre-Generation Guard for AI Video — August 2026
• I packaged the dark UI system I use for every client dashboard — August 2026
• The One Variable Rule That Fixed My AI Video Prompts — August 2026
• The Render Passed. The Video Still Wasn't Ready. — August 2026
• AIVideoRunbook: The Logbook That Turns Three Quality Gates Into One Pipeline — September 2026 · landing page
• AIVideoBatchQueue: The Batch Queue That Keeps Your 30-Clip AI Video Run Alive Overnight — September 2026
• ComfyUI-VideoGate: The QC Gate That Stops Broken AI Video Before It Ships — September 2026
• Every AI Video Delivery Comes With a Drawing — August 2026
• I turned a Claude Co-work tutorial into a complete motion graphics studio — July 2026
• AI Video Is 80% There. Here's the Last 20% — July 2026
• Video Workflows That Actually Ship — July 2026
• I turned a Remotion masterclass into a kit you can run — July 2026
• The Character Kit That Stopped Face Drift in My Client AI Videos — July 2026
• I built a character bible for AI video that survives past shot three — August 2026
• How I Built My Portfolio Site, Start to Finish — July 2026
• How I Actually Choose AI Models for Motion Graphics — July 2026
• Curated Prompts for Cinematic AI Video — July 2026
• n8n Workflows for AI Video Studios — July 2026