Projects.
©2026Four ventures under one umbrella. Each is real — in production or in pilot — and each links to the next.
Design & web — the studio itself.
Seam Studio is the design and engineering practice behind the other three ventures — and the same work is open to clients. Websites, brands, and the automation that runs underneath them, in Turkish and English, from first sketch to production.
One accountable pair of hands designs the interface and wires the system behind it, so the site and its machinery ship as one piece — not a front end handed off to someone else to plug in.
- Marketing sites, portfolios and stores, designed and built end to end
- Art direction, UX/UI, and a brand system that stays consistent everywhere
- Built in Framer or Next.js, with a CMS the client can actually run
- Bilingual Turkish / English copy and layout by default
- The workflows behind the site — reservations, inboxes, quoting, reporting
- n8n orchestration with LLM agents and third-party integrations
- Email and WhatsApp automation, always behind a human approval gate
AI defect detection for textile production lines.
Computer vision that inspects fabric and flags defects as the line runs. A YOLOv8-class model returns a bounding box, defect class and confidence score in under two seconds — and a built-in AI analyst turns those detections into root-cause insight for the QC team. Built on Next.js and Supabase, with a live demo you can try on your own fabric.
- Holes, stains and streaks; weft and warp defects
- Needle marks, broken stitches, pinched fabric, knots & slubs
- Tear/cut in training — new custom classes added without code changes
- Every hit returns a bounding box, class and confidence score
- Single-image or batch processing of hundreds of photos at once
- 14-day quality trends, defect distribution and live pass-rate
- A per-line AI analyst that reads the full dataset and recommends fixes
- Searchable, timestamped archive of original and annotated images
- Unlimited production lines with comparative benchmarking
- Audit-ready evidence, PDF / Excel export and supplier scorecards
- One-click confirm or draw-to-correct — corrections become training data and queue the model for retraining
In use on real production lines at KoruSer and Hassan Textile. Starter, Pro and Enterprise tiers, currently in early access.
RFQ email in, finished quote out.
For made-to-order manufacturers buried in quotation requests. Quotewright reads an incoming RFQ — the whole thread, not just the last message — pulls out every line, matches it to your own catalogue, and drafts a complete branded quotation in about two minutes, in the customer’s language. Then it stops: nothing sends without you.
- Classifies incoming RFQ emails and analyses the full thread
- Extracts size, colour, packaging and quantity, line by line
- Handles requests in multiple languages — TR, EN, DE, BG, FR
- Matches every line to your own catalogue, box math included
- Cash and term pricing on each line, written in the customer’s language
- Multi-currency with daily central-bank FX; EXW, FOB, CIF incoterms
- Prices come only from your catalogue — it will not invent a number
- Incomplete lines are flagged with exactly what’s missing; uncatalogued items route to sales
- One-tap approval from your phone — nothing sends without you
Built on real factory data — messy cost workbooks, multi-message threads, half-specified requests — and validated against a manufacturer’s historical quotes to the cent.
An AI maître for restaurants & hotels.
Misafir — Turkish for “guest” — watches every channel a venue is reachable on, drafts a reply in the guest’s own language, and runs the day-to-day admin around it. It handles the noise; the owner keeps the judgement. Nothing sends until you approve it.
- Google Reviews, Booking.com, Tripadvisor and Google Business profiles
- WhatsApp Business, Instagram and email — triaged across every venue
- Drafts in the guest’s language: TR, EN, DE, IT, FR, EL, matched to your venue’s tone
- Reservation confirmations and day-before reminders
- No-show follow-ups, waitlist, and dietary / allergy questions
- Post-visit thank-you with a review request; concierge asks — late checkout, taxi, recommendations
- Event and private-dining enquiries flagged for you
- Instagram captions and content calendars
- Weekly owner briefings comparing performance across venues
- Competitor monitoring and daily staff briefings
After-hours coverage so no enquiry sits cold overnight. Learns from your edits, GDPR & KVKK aware, and offered in three tiers — Front Desk, Front of House and Maison.