Frequently asked questions

FORGE OS is an AI-native procurement execution engine. You give it a product brief — what you want to make, in plain language — and it discovers candidate manufacturers, qualifies them, generates professional RFQs (requests for quotation), sends them from your own email, and reads the quotes that come back so you can compare them side by side. It replaces the spreadsheet-plus-guesswork phase of sourcing with one structured workspace.

Yes — that's precisely who it's built for. You describe your product the way you'd describe it to a friend; FORGE handles the industry mechanics: which suppliers to approach, what a proper RFQ must contain, what MOQ and lead-time answers mean, and how to read a quote. The workflow was designed by people who have sourced professionally since 2012, so the 'what do I even ask a factory?' problem is solved for you.

Nothing you couldn't do — it does it in minutes instead of days. A sourcing run performs the discovery, legitimacy screening, and contract-manufacturer qualification you'd normally spread across directories, LinkedIn, and reference calls, then drafts send-ready RFQs and normalizes incoming quotes into one comparison table. You keep every judgment call; FORGE removes the grunt work and gives you a wider, better-screened funnel per brief.

No. FORGE doesn't sell listings and suppliers can't pay to appear. Candidates are generated per brief by AI trained on the global manufacturing landscape, enriched with live web evidence, and scored for legitimacy. Every run is specific to your product, your materials, and your constraints — not a static database dump.

Every candidate is scored on factory confidence and intermediary probability, using signals like export records, registrations, and web evidence. Traders and brokers are flagged as such. FORGE also applies a stricter test most tools skip entirely: does this company manufacture for third parties at all? Brand-owners who run factories only for their own products are automatically disqualified from outreach — a world-class factory you can't hire is not a supplier.

Your briefs, assets, and quotes live in your private workspace, isolated per account with row-level security. Nothing you upload is shared with suppliers until you explicitly send an RFQ — and every send passes a human confirmation gate. Uploaded assets are analyzed to grade their sensitivity, so you know what should stay behind an NDA before anything goes out. We don't sell or share your data. See our Privacy Policy for the specifics.

Never. RFQs are drafted by AI but dispatched only after you review and confirm each one, and they send from your own connected Gmail account — suppliers see you, not a platform. Incoming replies are read and parsed into structured quotes, and quote data is only written to your records after a review step. Human gates sit on every irreversible action.

One sourcing run takes one product brief through discovery, qualification, and ranking — typically producing a ranked slate of candidate manufacturers with scores and reasoning. One run per product is usually enough to start outreach; teams sourcing multiple products or revisiting categories each month simply pick a plan with more runs. Plans range from Explore (3 runs/month) to Enterprise (unlimited).

Consumer products across the board: apparel and accessories, footwear, sporting goods, beauty and personal care, home and lifestyle, packaging, electronics and hardware, promotional and custom manufacturing. The team behind FORGE has developed products in most of these categories since 2012, and the qualification logic is category-aware — a padel racket brief and a skincare brief get very different screening.

The discovery engine is global. In practice most consumer-product runs surface manufacturers across China and wider Asia, plus Europe, North America, and Latin America, depending on the category and your constraints — you can steer regions in the brief. Ateliersavant's own vendor experience spans Los Angeles, Mexico, China, France, the UK, and Japan, and that operating knowledge shaped the engine.

A good agent brings judgment and relationships — and costs a retainer or a percentage, works one category at a time, and takes weeks. FORGE gives you the structured funnel an agent would build (discovery, vetting, RFQs, quote comparison) in minutes, for a flat monthly price, with every step visible and yours to control. Many users do both: run FORGE first, then bring an agent or the Ateliersavant team in for negotiation and production management.

Marketplace search shows you who paid to be found and leaves the vetting, outreach, and quote-wrangling to you. FORGE generates candidates independently of ad placement, screens out brokers and captive brand-owners, writes and sends professional RFQs, and turns messy quote emails into a clean comparison. It's the difference between a search box and an execution engine.

Yes. When suppliers reply to your RFQ, FORGE ingests the thread, extracts pricing, MOQs, lead times, and terms with AI, and lines every quote up in a comparison table. Extraction runs behind a review gate — you approve what gets written to your records — and a rule-based fallback catches formats the AI is unsure about.

A product brief — even a rough one — and a plan. Everything runs in the browser; there's nothing to install and no API keys to manage. Connecting Gmail (for sending RFQs from your own address) and Google Sheets (for exports) are optional one-click steps you can do when you need them.

Yes — plans are monthly with no lock-in, managed through Stripe's secure billing portal. If you cancel, your workspace data remains yours; you can export your supplier and quote records, and you can request full deletion at any time in line with our Privacy Policy (we're GDPR-conscious by design — the company behind FORGE is French).

FORGE OS is built by Ateliersavant, a product development and sourcing practice operating since 2012 — hundreds of products developed across nearly every consumer category, with manufacturing partners from Los Angeles to Tokyo. FORGE is the systematization of that practice: the same discovery, vetting, and RFQ discipline used on real client work, encoded into software. Read the full story on our About page.