Why Google Sheets is not at all the right UI/UX for this system
1) This is a “studio workflow,” not a spreadsheet workflow
The requirement is an animation-style pipeline: script → shot list → storyboard
layout → final images → video → approvals and regen loops
A spreadsheet UI will feel like an internal hack, not a production tool—especially
once you add:
reference uploads (characters/backgrounds/style)
visual review panels (thumbnails, zoom, side-by-side comparisons)
per-shot regen notes + attachments
versioning (v01/v02)
video preview & approval
Sheets can’t do those experiences cleanly without becoming messy links
everywhere.
2) Sheets will break “human-in-the-loop” in subtle ways
Your process is gated: “nothing moves forward without approval,” and per-stage
revisions must be controlled
. In Sheets:
comments get scattered
approvals are inconsistent (people type “ok”, “yes”, etc.)
attachments are hard to validate reliably
multi-approver flows become fragile
You need a UI that enforces structured actions:
✅ Approve / 🔁 Regenerate / ✍️Edit / 📎 Attach / 🧪 Compare
3) If this ever becomes external-facing, Sheets becomes a dead-end
Even if this is internal for now, our vision is a pipeline that can scale into a tool.
Sheets won’t support:
clean access control per user/project
client-friendly UX
audit trails & role-based approvals
long-term maintainability
What we should build instead (intended architecture)
Your intended approach is:
✅ Webflow = The front-end experience (I saw that you already have
experience in that)
Users interact with:
guided steps
upload fields
review screens with thumbnails/video players
approval buttons
regen notes forms
✅ n8n = The orchestration brain
n8n does:
calls LLMs / NanoBanana / Veo 3.1 APIs
stores files to GCS
updates DB status
triggers next step only after approval
✅ GCS = File storage
Everything file-based lives here:
script versions
[Link]
layout storyboard frames
final images + overlays
final videos + previews
✅ Database = System memory & logs
DB stores:
project/scene/shots
statuses + approvals
links to GCS objects
generation logs and regen notes
This is exactly aligned with what your requirement describes: staged outputs,
editable loops, approvals, and scale-ready storage
REGARDING API ACCESS:
Thanks for outlining what you need from our side. We will definitely provide the
required API access (NanoBanana Pro, Veo 3.1, Claude, etc.), but we’d like to
introduce those a little later in the process. You can begin development
immediately without them by implementing the system in what we call a Mock
Mode first.
This is actually the standard approach for building production pipelines like this.
Why live APIs are not required at this stage
Most of the work right now is workflow orchestration and UI behaviour, not
model execution.
Specifically, you can already build and test:
Webflow interface (forms, step-by-step flow, review/approval pages)
Webflow → webhook → n8n triggers
Stage transitions (script → review → lock → next phase)
File storage logic and folder structure
Versioning (v01, v02) and regeneration handling
Job tracking, status updates, and logs
Partial reruns (e.g., regenerate one shot only)
All of that can be validated using placeholder responses instead of real model
calls.
Phase 1 (Scriptwriting) can also be developed without our accounts
For the initial script-development phase, where we use LLMs (Claude / GPT /
Gemini) to generate:
logline
synopsis
beat sheet
scene outline
draft script
analysis/refinement
you can use your own API access or even free-tier accounts for testing.
We are not tied to a specific account during development — what matters is:
defining the prompt structure properly,
ensuring responses are captured and versioned correctly,
validating the review/edit/regenerate loop.
Later, we will switch the credentials to our managed accounts without changing
the workflow logic.
Same applies to Google services during build
For GCS / Google integrations, you can start with:
your own Google Cloud project (free tier is fine),
test buckets for upload/download validation.
Once the pipeline is stable, we will replace those with our production
environment and credentials.
This lets you proceed immediately while we finalize access provisioning on our
side.
When we will provide production keys
We will share required credentials asap, but don’t see these as blockers to your
work and progress at all! We can even share these once:
the orchestration flow is complete,
data structures and storage paths are confirmed,
we are ready to test real outputs instead of placeholders.
This approach avoids unnecessary usage costs, keeps development fast, and
ensures credentials are added only when the system is ready.
We can discuss any queries you may have in the meeting.
Regarding N8N access:
You can build the workflows on your own n8n instance for now.
Once the pipeline is validated, we will migrate everything into our environment
and replace credentials with ours.
Please keep the implementation portable (no hardcoded keys, configurable
endpoints), so the transfer is straightforward.