FashionBot — Clawd Trade Build Plan

The Product in One Sentence

v1.0

FashionBot is a WhatsApp business number that gives any fashion designer access to Fashion Forge's full AI + expert network — tech packs, patterns, sketches, sourcing, fitting coordination — all via text and voice notes, no app, no login, no minimum.

📱

WhatsApp Native

Works on any phone, no download

⚡

AI-First

60–85% automated; human for the rest

◎

Expert Validated

Artiquity Principle on every output

Tech Pack Conversation Flow

Wave 1

I need a tech pack for a double-breasted wool blazer. Women's XS–XL. Nordstrom contemporary. LA made.

FashionBot → intent: TECH_PACK · routing to Forge API

Got it. A few quick questions to get this right:
1. Any lining? (Full / half / none)
2. Peak or notch lapels?
3. Target MSRP?

Or voice note me the details — I'll handle the rest.

Full Bemberg lining, peak lapels, $495 MSRP. Double welt pockets, no vents.

Forge API → generating tech pack + pattern intel Gemini

✓ Tech pack generated.

14 finished measurements (XS–XL), full BOM (shell, lining, interlining, buttons, thread), stitching specs, 6 QC checkpoints.

Assigned to Elena R. — LA pattern maker, tailoring specialist. Validation est. 48 hrs.

📄 Download: fashion-forge.html
📐 Pattern intel: included
💰 Est. cost-to-make: $89–$112

Can you also find me a Italian wool crepe source? 380gsm+

intent: SOURCING → routing to Sourcing agent BYO API

3 verified Italian suppliers matched your spec:
1. Lanificio Cerruti — 380gsm wool crepe, $28/yd, MOQ 30m
2. Vitale Barberis Canonico — 400gsm, $34/yd, MOQ 20m
3. Loro Piana — 395gsm, $52/yd, MOQ 10m

Want swatches coordinated? Human sourcer

Pattern Maker Matching Flow

Wave 2

I need someone who can drape a bias-cut silk charmeuse dress. Very complex. I've had bad luck with CLO3D — I need a real draper.

intent: PATTERN_MATCH · garment_type: bias_cut_silk · skill: draping

Draping is our least-automated service (25% AI) — exactly right to want a human.

We have 3 drapers in our LA network who specialize in silk charmeuse bias cuts:

🧵 Carmen V. — 18 yrs experience, Azzedine Alaïa trained. 2 week lead. $850/style.
🧵 Liu W. — 12 yrs, bias cut specialist. 1 week lead. $700/style.
🧵 Roberto M. — 20+ yrs, evening wear focus. 3 week lead. $950/style.

Which direction?

Carmen. Let's go.

booking → calendar + deposit link Human handoff

✓ Carmen V. booked.

Deposit link (50%): [pay link]
First fitting: Tue Mar 10, 10am LA
Address: 824 S Los Angeles St

I'll coordinate the fabric delivery, prep your design brief, and send Carmen your spec sheet. You just show up.

System Architecture

Build Spec

Interface Layer

Twilio WhatsApp Business API Voice note → Deepgram STT Fal CDN → WhatsApp media PDF download links (GitHub Pages)

↓

Intent Router

TECH_PACK FIRST_PATTERN PRODUCTION_PATTERN DRAPING SKETCH SOURCING FITTING_BOOK PATTERN_MATCH DESIGN_DEV BID_ESTIMATE

↓

Forge API (this repo → Cloud Run)

Gemini Pro (tech pack gen) Fal flux (technical sketches) Pattern intelligence engine Grading calculator BOM generator Sourcing matcher Cost estimator

↓

Human Network Layer (Artiquity Principle)

Pattern Maker directory (LA, NYC) Specialization tags (bias, tailoring, denim...) Availability + booking calendar Validation queue + notification Fit model roster Bill of Rights enforcement

↓

Data Layer

PostgreSQL (users, brands, jobs, payments) Brand fit DNA store (per-client) Pattern library (versioned) Supplier database Knowledge capture (video, annotations) Stripe billing

Bot Commands — What Designers Say, What Happens

Trigger / Intent	What the Bot Does	Wave
"tech pack for [garment]"	Runs Forge API → generates measurements, BOM, stitching, QC, care, labels → sends PDF link + assigns pattern maker for validation	Wave 1
"sketch / flat for [garment]"	Generates technical flat (front + back) via Fal flux → sends images in WhatsApp + download link	Wave 1
"first pattern for [garment]"	Generates pattern pieces, grading table, ease allowances, cutting instructions → assigns master for tacit-knowledge validation	Wave 1
"find me a pattern maker for [spec]"	Matches from human network directory → returns 3 options ranked by specialization + lead time + price → books on confirm	Wave 2
"production pattern"	Converts approved first pattern to full grade range, marker optimization, factory output formats (Gerber/Lectra/DXF)	Wave 2
"source [fabric spec]"	Searches supplier database → returns 3 verified options with price/MOQ/lead time → coordinates swatch request	Wave 2
"draping for [garment]"	Routes to human draper directory — cannot be significantly AI-automated. Returns human availability + pricing + booking.	Wave 2
"design dev / development"	Trade-off analysis: cost-to-make vs. design intent, simplification options, BOM cost breakdown, pricing architecture	Wave 2
"book a fitting"	Coordinates fit model from roster → schedules → sends address + deposit + day-of brief	Wave 3
"estimate cost" / "what would this cost"	Returns materials + labor + trim + permit cost estimate for LA manufacturing based on garment type + fabric + size run	Wave 1
"remember:" / brand notes	Stores brand fit DNA, preferred pattern makers, colorway preferences, factory contacts in PostgreSQL user record	Wave 1
Voice note	Deepgram STT → intent detection → same flow as text commands. "Voice note me the details — I'll handle the rest."	Wave 2

Build Milestones

8-Week Sprint

Week 1–2: Core Infrastructure

Wave 1 Foundation

Twilio WhatsApp Business setup · Webhook handler (FastAPI) · Intent router (Gemini classifier) · Session/memory (PostgreSQL) · Brand setup flow

Deliverable: Bot responds to "tech pack for X" with a generated document

Week 3–4: Forge API Integration

Wave 1 Core

Connect to fashion-forge.html backend · Tech pack JSON → WhatsApp-formatted summary + PDF link · Fal sketch generation + image delivery · Pattern intelligence output + human assignment queue

Deliverable: End-to-end: WhatsApp in → tech pack PDF + sketch images out

Week 5–6: Human Network Layer

Wave 2 Launch

Pattern maker directory (seed with 5–8 LA practitioners) · Specialization tags + availability · Booking flow → calendar + Stripe deposit · Artiquity credit system (expert name on every output)

Deliverable: Bot matches and books a real pattern maker in conversation

Week 7–8: Billing + Brand DNA

Revenue-Ready

Stripe per-project billing (tech pack $299, pattern $399, sketch $99) · "Remember:" brand fit DNA storage · 14-day free trial → subscription tiers · First paying client onboarded

Deliverable: Revenue flowing. FashionBot live for beachhead designers.

Pricing Architecture

Per-project + subscription tiers

Starter

Free

14-day trial

3 tech pack generations

3 sketch generations

Basic pattern intel

No human validation

Studio

$149

/month

Unlimited tech packs + sketches

5 expert validations/mo

Brand fit DNA storage

Pattern maker matching

Voice note support

Forge Pro

$399

/month

Everything in Studio

Unlimited expert validations

Priority pattern maker booking

Fitting coordination

Sourcing + cost estimation

Production sampling oversight

À la carte (per project)

Tech Pack (per style)$299

First Pattern (per style)$399

Technical Sketch (per style)$99

Pattern Maker Placement (staffing)20–25% of Y1 salary

Draping (human, per style)$700–$950

Tech Stack

Extends existing Clawd Trade infrastructure

Interface

Twilio WhatsApp Business Deepgram Nova-3 (STT) ElevenLabs (optional TTS) GitHub Pages (PDF delivery)

AI / Generation

Gemini Pro (intent + generation) Fal flux/schnell (sketches) Awareness Proxy (all AI calls)

Backend

FastAPI (Python) Cloud Run (GCP) PostgreSQL (Cloud SQL) Redis (session cache)

Billing / Ops

Stripe (subscriptions + per-project) Cal.com (booking calendar) Notion (pattern maker directory admin)

Reuses existing Clawd Trade infrastructure: The FashionBot API follows the same FastAPI + Twilio + PostgreSQL + Cloud Run pattern as `clawd-trade-api` in Revilopark. New bot file `fashionbot.py` + Forge API integration. Greg deploys with a new Twilio WhatsApp number + Stripe price ID. Same deployment script as RoofBot/PlumbBot.

What's Already Built vs. What Needs Building

✓ Already Live

✓ fashion-forge.html — Tech pack + pattern + sketch generator

✓ Awareness proxy — Gemini + Fal routes live

✓ Clawd Trade API infrastructure (FastAPI + Cloud Run pattern)

✓ FashionBot in clawd-personal.html (persona + system prompt)

✓ BYO proxy routes (BackyardOne integration)

✓ Dashboard (fashion-forge-dashboard.html)

○ Needs Building

○ fashionbot.py — Clawd Trade bot file with Forge API calls

○ Intent router — 10 fashion-specific intent classifiers

○ Forge API backend — FastAPI wrapper around the fashion-forge logic

○ Pattern maker directory — seed with 5–8 LA practitioners

○ Booking + payment flow (Cal.com + Stripe)

○ Greg: new Twilio WhatsApp number for FashionBot

○ Greg: Cloud Run deploy + PostgreSQL + Stripe setup

Build Plan:FashionBot as Fashion Forge

Build Plan:
FashionBot as Fashion Forge