Home / Video / Assets, Tooling, and Strategy
Affirmology: Website + Video Strategy Brief v1
Updated Jun 12, 2026 · Affirmology_WebVideo_Brief_v1.md
Summary. Date: June 12, 2026. Based on your folder (Demo Launch Plan, Brief v7 brand system, Exec Summary v3, Investor Brief, Lily trailer VO) plus fresh web research.
Affirmology: Website + Video Strategy Brief v1
Date: June 12, 2026. Based on your folder (Demo Launch Plan, Brief v7 brand system, Exec Summary v3, Investor Brief, Lily trailer VO) plus fresh web research.
Where you already stand (stronger than you think)
- Brand system is locked. Cream
#fbfaf7, gold #b48b3e, Playfair Display / Cormorant headlines, Inter body, JetBrains Mono accents. Every new asset inherits this. No design work needed, only execution.
- The 120s demo-page video has its soundtrack.
Affirmology_TrailerVO_Lily_slower_v1.mp3 runs 117 seconds. The tightPauses version runs 109s. The video is a visuals problem, not a script problem.
- The demo site plan and scaffolding exist. Cloudflare Pages + Worker + Mac mini API are already architected. Polishing the demo pages is a styling and animation pass, not a rebuild.
- The 6-min investor video has source material. Pitch deck, Investor Brief v1, Exec Summary v3 give the full narrative arc. It needs a video script, a shot list, and generated footage.
Higgsfield AI: yes, it is the right call
Higgsfield is a cinematography-first platform that bundles 15+ video models (Veo 3.1, Sora 2, Kling 3.0, Seedance, etc.) under one subscription, then adds its own control layer on top. Three reasons it fits Affirmology:
- Cinema Studio 3.5 gives 70+ deterministic camera presets (slow dolly, crane, 360 orbit, breathing zoom). Luxury-mystical brands live and die on camera movement. No competitor has an equivalent control layer.
- One subscription, every frontier model. You can A/B the same shot across Veo 3.1 and Kling 3.0 without juggling accounts. Sora's standalone app is being wound down by OpenAI, but the model stays accessible inside Higgsfield.
- Soul ID keeps a recurring character consistent across shots if you ever want an on-screen guide persona.
Plan to buy: Ultra tier (~$84 to $99/mo annual, 3,000+ credits). You chose "whatever it takes," and premium models burn 40 to 70 credits per generation. The 6-min video will need 60 to 120 generations including rejects. Plus tier (1,000 credits) would run dry mid-project. Buy top-up packs (~$5/100 credits) as needed; they expire in 90 days, so do not stockpile.
What Higgsfield is NOT: an editor. It generates clips. You assemble elsewhere (see workflow below).
Where it sits vs alternatives: Veo 3.1 + Google Flow is the best single-vendor path if you want native audio and SceneBuilder assembly in one place. Runway Gen-4.5 is the pro editor favorite. Kling 3.0 is the price/quality winner. HeyGen is only needed if you want a photoreal talking presenter; with Lily VO you do not.
Your learning path (you drive, ordered, ~4 hours total)
- Official Higgsfield tutorial playlist (start here, ~1 hr): "Higgsfield AI Tutorials: Start Here" by @HiggsfieldAI: https://www.youtube.com/playlist?list=PLaQe6ms8V2M4-XeOfKo3aZbov0VRsiKiU
- "The ONLY Higgsfield AI Tutorial You Need 2026: Step-by-Step for Beginners" (~30 min): https://www.youtube.com/watch?v=R7GZjRMsrzM
- "Higgsfield AI Full Tutorial for Beginners 2026 || Complete Guide": https://www.youtube.com/watch?v=CVNbDUiqqn4
- Optional, if you also try the Google path: "Google Flow Tutorial 2026 (Complete Step-By-Step Guide)": https://www.youtube.com/watch?v=RbvpUAnyZbc and Google's "5 tips for using Flow": https://blog.google/innovation-and-ai/products/flow-video-tips/
After tutorial 1, generate five test shots from the shot list I will write you. That is the fastest way to learn: real shots for the real project, not toy prompts.
Workflow: the two videos
Video A: 120s demo-page trailer (first, smaller, teaches you the pipeline)
- I write a beat-by-beat shot list timed against the 117s Lily VO (roughly 18 to 24 shots of 4 to 8s each), each with a ready-to-paste Higgsfield prompt and camera preset in the cream/gold visual language.
- You generate in Higgsfield. Expect 2 to 3 takes per shot. Keep a simple accept/reject folder.
- Assemble: clips + Lily VO + music bed. Easiest tools: CapCut (free, fast) or DaVinci Resolve (free, more control). Either is a one-evening learn; the edit itself is mostly laying clips against an existing VO timeline.
- Export 1080p (or 4K), drop on the demo page as the hero or as a "watch first" gate.
Video B: 6-min investor/education video (after A, reuses everything you learned)
- I draft the script from your pitch deck + Investor Brief + Exec Summary: problem, the four engines, the demo experience, market, ask. You edit for voice.
- Record VO in ElevenLabs with your chosen voice (you already have the account and auditioned voices).
- I produce the shot list (~50 to 70 shots, mixing Higgsfield b-roll with screen recordings of the actual demo, which investors trust more than pure AI footage). Real product on screen is the credibility anchor; AI footage is the emotional wrapper.
- You generate, we assemble the same way as Video A. Budget 2 to 3 weeks part-time.
Demo page polish (I build this, you review)
Claude-built static pages with GSAP ScrollTrigger animation, deployed on your existing Cloudflare Pages. Free hosting, no lock-in, total brand control. Framer ($30/mo) stays the fallback if you ever want drag-and-drop, but you do not need it.
Skills worth installing to raise the design ceiling (all free, open source):
- gsap-scrolltrigger skill (production-grade scroll animation recipes): https://github.com/freshtechbro/claudedesignskills
- claude-frontend-skills (anti-generic, distinctive design direction): https://github.com/Koomook/claude-frontend-skills
- Frontend Design Toolkit (curated skills/MCPs for better-looking frontends): https://github.com/wilwaldon/Claude-Code-Frontend-Design-Toolkit
- Already in your Cowork:
web-artifacts-builder. Curated indexes if you want to browse: https://github.com/travisvn/awesome-claude-skills
Build scope per your Demo Launch Plan: intake page + results page in Brief v7 language, breathing gold ✦ progress state, marked.js report render, sticky audio bar, plus a 120s video slot.
Action plan
This week
1. You: subscribe to Higgsfield Ultra, watch the official playlist (1 hr).
2. Me: write the 120s trailer shot list with paste-ready prompts, timed to Lily VO.
3. Me: build the polished demo intake + results pages (mocked backend per Demo Launch Plan step 1).
Next week
4. You: generate trailer shots, assemble in CapCut or Resolve.
5. Me: draft the 6-min investor video script from your existing docs.
Weeks 3 to 4
6. 6-min video: VO, shot list, generation, assembly. Demo pages go live on affirmology.ai/demo.
Sources
Higgsfield capabilities and pricing: https://higgsfield.ai/ , https://www.imagine.art/blogs/higgsfield-ai-pricing , https://filmora.wondershare.com/ai-generation/higgsfield-ai-review.html , https://aifunnelinsider.com/higgsfield-ai-review-2026/
Tool comparisons: https://tooldirectory.ai/blog/best-ai-video-generator-2026-veo-runway-kling-pika-luma , https://lushbinary.com/blog/ai-video-generation-sora-veo-kling-seedance-comparison/ , https://www.colossyan.com/posts/heygen-vs-synthesia/
Website workflow: https://www.mindstudio.ai/blog/animated-3d-websites-claude-code-ai-video-generation , https://www.builder.io/blog/3d-gsap , https://goodspeed.studio/blog/framer-review