AI Fashion Stylist
Personal styling and wardrobe agents grounded in real garment intelligence.
Akwa Platform
Akwa is building the structured design layer for fashion: from sketches and semantic garment graphs to production-oriented technical data. The first door is open: on a paid plan, a personal API key connects your AI assistant or your own system to the garment graph, tailor briefs, Design Trust, exports and Tech Packs. Early access, built on something real, not a promise.
Keys are created in Settings > Developer access. Partner capabilities beyond the first door open through the waitlist.
Connect your assistant Join the developer waitlistThe data model
Most fashion AI ends with an image. Akwa treats a design as data every stage can read and enrich: seams, panels, points of measurement, provenance. The semantic garment graph is the product; an API is just how you reach it.
Why fashion needs structured design data
Software has code. Architecture has BIM. Film has the edit timeline. Music has the score. Each is a structured source of truth that every tool can read and every stage can build on without guessing.
Fashion has an image and a PDF. A render is pixels; a tech pack reverse-engineers those pixels; a factory re-interprets the tech pack. Design intent leaks at every handoff, and the cost shows up as failed samples, poor fit, misread details and slow production.
When a garment is structured data instead (seams, panels, points of measurement, construction and provenance) every stage reads and enriches the same source rather than reinterpreting a picture. That programmable design layer is what fashion is missing, and what Akwa is building.
What you could build
Personal styling and wardrobe agents grounded in real garment intelligence.
Turn a design into production documentation inside your own product.
Measurement-based fit guidance for commerce and styling, on a declared Body Twin, no photographs.
Brief-to-garment guidance for tailors, ateliers and made-to-order.
Complexity, readiness and construction intelligence for sampling.
Teach construction, culture and craft with structured fashion knowledge.
Let assistants design, compare, visualise and buy, on the user's behalf.
Capabilities
Culturally grounded fashion concepts from language, sketches or structured briefs.
Structured sample-stage tech packs: flats, BOM, specs, construction and QC. Requestable through the API.
Originality signals, cultural provenance, similarity risk and readiness. Readable through the API and MCP.
A measurement-based representation of the wearer: declared attributes, no photographs. Try-on is not offered today, and the Body Twin is not yet readable through the API.
Connect Claude Code, VS Code or Copilot to your Akwa account with a personal API key: garment graphs, tailor briefs, Design Trust and exports.
Production complexity, construction risk and factory-readiness estimates.
Connect
On a paid plan, create a personal API key in Settings > Developer access. It opens the Akwa MCP server and the Akwa API. Both answer only about your own designs and the published Akwa house collection, never with renders, and every value carries its provenance state (asserted, inferred, unstated). Send it as Authorization: Bearer akwa_sk_….
Early access · paid plans · scopes: read, export, generate · keys revocable at any time
Streamable HTTP, JSON-RPC 2.0 over POST. Five tools: get_garment_graph, get_tailor_brief, run_design_trust, list_unresolved_confirmations (read-only) and export_design (a layered SVG for Illustrator, or JSON, one-hour link).
https://olgggkmqnncdjdxwikml.supabase.co/functions/v1/sgg-mcp
One command, then ask Claude about your designs by name or AKW identity.
claude mcp add --transport http akwa https://olgggkmqnncdjdxwikml.supabase.co/functions/v1/sgg-mcp --header "Authorization: Bearer akwa_sk_..."
Add to .vscode/mcp.json (or copilot mcp add with the same URL and header).
{ "servers": { "akwa": { "type": "http", "url": "https://olgggkmqnncdjdxwikml.supabase.co/functions/v1/sgg-mcp",
"headers": { "Authorization": "Bearer akwa_sk_..." } } } }The same reads by design id or permanent garment identity, exports, and Tech Pack requests (EUR 25 plus a fixed surcharge, or included in your plan). GET /v1 lists every route.
curl -H "Authorization: Bearer akwa_sk_..." https://olgggkmqnncdjdxwikml.supabase.co/functions/v1/api-v1/v1/designs
Consumer connectors sign you in with OAuth 2.1 (early access). Add a custom connector with the MCP URL below, approve it with your Akwa account, and disconnect it any time from Settings. Akwa inside ChatGPT is entitlement-only: what your plan already includes, with no purchases in the chat.
https://www.akwa.design/mcp
Platform philosophy
AI assists; designers decide. Faithful by default, creative only by request.
A garment is data first and an image second. The graph, not the render, is authoritative.
Origin is the starting material, grounded in a cultural fashion ontology, not a surface filter.
Everything resolves toward something a real tailor or factory can actually make.
The missing layer
Akwa is building the operating system behind AI fashion development: not another chatbot, but reliable, programmable fashion capabilities. Design intelligence, a cultural ontology, manufacturing knowledge, technical documentation and privacy-first digital twins, all reachable from one platform.
Early access
The waitlist shapes the roadmap: what you want to build tells us which capability to open first. No spam; the founder reads every one.
Claude and Claude Code are trade marks of Anthropic, PBC. ChatGPT is a trade mark of OpenAI. Visual Studio Code and GitHub Copilot are trade marks of Microsoft Corporation. Adobe and Illustrator are trade marks of Adobe Inc. Akwa is not affiliated with, endorsed by or sponsored by any of them. Names are used only to identify the tools a key connects to.