Vendo

Selling to businesses? "Catalogs" in Spree Commerce let you set which products each business customer can buy and what they pay. It's just one of many open-source B2B features. 1️⃣ One catalog is one agreement. Assign one customer to it, or many. 2️⃣ Divisions inherit their parent company's catalog. No extra setup. 3️⃣ The catalog decides which products that customer can buy. 4️⃣ And what they pay: per product, one percentage, or a price per quantity. Set it up once. Sell to businesses the way business customers buy. Try it for free: npx create-spree-app@latest my-store #ecommerce #opensource #B2B #wholesale #distribution

Spree Commerce 5.6 lets you gate prices and checkout behind a login, hiding them from guests until they request an account: spreecommerce.org/announcing… Included in this release is a Next.js storefront with a wholesale portal running alongside the retail shopping experience. It's just one of many uses cases of price & checkout gating. 👉 Same storefront. 👉 One backend. 🖐 Multiple customer segments with their own pricing. Here is the part that matters if you have built this before: the gating is enforced in the API, not in the storefront code. Prices stay hidden whether the request comes from the Next.js storefront, a mobile app, or a custom frontend. A new frontend cannot leak a price you meant to keep private. Also included in this release: ↗️ Gated sales channels - require login, hide prices, or disable guest checkout, enforced at the API layer ↗️ Wholesale portal - a Next.js storefront app with customer pricing and quick ordering ↗️ Localized admin - admin panel in 50-plus languages, per team member, with right-to-left layouts ↗️ Pre-orders - sell before stock arrives, set per variant with an expected ship date ↗️ Simpler deployment - one Docker image, no required Redis, 40% smaller than 5.5 ↗️ Redesigned emails - localized, mobile-friendly transactional emails Demo: demo.spreecommerce.org/us/en Try it out for yourself with a free Sandbox: a cloud test store ready seconds after you sign up. No installation. No credit card. Nothing to cancel. Try it for free in your browser: spree.sh #wholesale #B2B #ecommerce #opensource #headlesscommerce #spreecommerce

Spree Commerce Sandboxes are here: free cloud test stores ready seconds after you sign up. No installation. No credit card. Nothing to cancel. More on this: spreecommerce.org/announcing… Developers could always run the code locally with one command: npx create-spree-app@latest my-store The sandbox opens the same evaluation to the whole team: sign up and click around a live store for yourself. What's inside every sandbox: 1️⃣ Full admin dashboard: products, orders, promotions. 2️⃣ Store API and Admin API for testing live with sample data. 3️⃣ Its own web address: share it or throw it away. 4️⃣ Payment gateways in test mode: demo-safe by default. Developers get a live backend to play around with Store and Admin API. Business teams get an admin dashboard they can explore without asking anyone to provision it. Agencies get a client demo that takes less time to create than a meeting invite. Try it for free in your browser: spree.sh/ #ecommerce #opensource #headlesscommerce #spreecommerce

Sales Channels in open-source Spree Commerce let you sell online, in your stores, and mobile app from one backend, so customers get the same prices, stock, and checkout everywhere they shop: spreecommerce.org/ecommerce-… Most retailers run each of those on its own system. Catalogs drift, prices fall out of sync, and no one report shows what each surface actually earns. Not to mention bad customer experience - no click & collect or cross-channel returns & refunds. Sales Channels put every surface inside one store. Every order is tagged with the channel it came from. So you see what your wholesale line earns next to your website, side by side, with no exporting and stitching. Each channel keeps its own catalog and schedule. A product can go live at the register today and online next month, set once. Buy online, pick up in store, return at the counter. It works, because underneath it is one system. One backend. Every surface. One clear view of the business. Free and open source. #ecommerce #opensource #retail #omnichannel

Your Spree Commerce backend is fully programmable. No need for an admin panel. Truly API-first. With an Admin API and a fully-typed TypeScript SDK, your code can read and write: 👠 products, 🛒 orders, 👱 customers, 🛍️ promotions, so the same things you manage in the admin. Connect any system and automate any workflow: 👉 Feed your catalog from a PIM or ERP. 👉 Sync orders into a warehouse system the moment they're placed. 👉 Build a custom admin inside your own SaaS instead of a second login. What makes it safe and fast at scale: 1️⃣ Scoped keys: each integration touches only the data you allow. 2️⃣ Bulk operations: changes run as single queries, so large catalogs don't grind. 3️⃣ Draft Orders: compose a complete order from an external system. 4️⃣ A fully-typed SDK: autocomplete for every endpoint. And because Spree Commerce is open source, both the storefront and backend APIs are in the free core, on infrastructure you own. Your systems. Your data. Your infrastructure. Check it out: spreecommerce.org/ecommerce-… Try Spree for free: npx create-spree-app@latest my-store #ecommerce #opensource #API #headless #developertools

Announcing Spree Commerce 5.5: faster AI delivery, better code quality, cheaper agentic development and automated daily commerce operations. The timing is no accident. The Stack Overflow 2025 Developer Survey found that 66% of developers say their biggest frustration are “AI solutions that are almost right, but not quite.” That frustration is exactly what this Spree Commerce open source release fixes: spreecommerce.org/announcing… Here is what you get: 1️⃣ Your AI agent becomes an eCommerce expert. 25 installable skills teach it how Spree Commerce works before it writes code. Works with Claude Code, Cursor, Copilot, Codex and others. 2️⃣ Token cost drops about 90%. Your agent stops burning tokens guessing at best practices and long-term solutions. 3️⃣ Automate your store back office. A new Admin API and TypeScript SDK sync orders, products, and customers with your ERP, warehouse, or custom dashboard. 4️⃣ Let your AI agent run the store for you with Admin API and CLI. Let it do all the daily chores like restock items, fulfill orders, and launch promotions. 5️⃣ And for complex catalogs: Sales Channels, Stock Reservations, and rules-based Order Routing. Built in, not plugins. Build faster, for less with Spree Commerce open source. Full announcement: spreecommerce.org/announcing… You own the code. Self host it anywhere. Control everything. Start free: npx create-spree-app@latest my-store #ecommerce #opensource #AI #headless #developertools

67% of B2B buyers prefer NOT to talk to a rep (Gartner, 2026). "Don't call me. I'll call you" they say. Why? To avoid repetitive cold outreach. Mostly automated. They would rather (re)order through a B2B website themselves. Here is the dynamic. A medical supply buyer reorders the same gloves and gauze every week. She stays with the supplier whose wholesale website shows her own contract price and lets her check out (quick-order) in three minutes. The distributors still relying on reps and running on faxed POs and emailed PDF price lists feel it first. The fix is an online store that knows each buyer: their negotiated price, their volume discounts, their reorder history. Not one public price shown to everyone. Spree Commerce is an open-source eCommerce platform built for exactly this. Its free Community Edition includes all four: 1️⃣ Price Lists: negotiated prices and quantity tiers per account. 2️⃣ Customer Groups: each buyer mapped to the right pricing. 3️⃣ Gated storefronts: contract terms stay behind a login. 4️⃣ Account order history: the weekly reorder in three clicks. When a hospital network needs roles, spending limits, and manager sign-off, the Enterprise Edition B2B module adds Buyer Organizations and Approval Workflows on the same platform. You host it yourself, so the code, the contracts, and the audit records stay yours and confidential. The distributors winning accounts right now are the ones whose buyers are enabled to self-serve instead of pestered with upsell messages and phone calls. More on self-serve wholesale portals: spreecommerce.org/medical-su… #ecommerce #opensource #B2Becommerce #medicalsupplies #healthcare

Spree Commerce now partners with @Adyen to let you accept payments in 150 currencies and a multitude of local payment methods. iDEAL in Amsterdam 🇳🇱 Bancontact in Brussels 🇧🇪 Same store. Same checkout. Different payment method. A card-only checkout loses European 🇪🇺 sales before the payment page loads. Why? a) iDEAL handled roughly 70% of Dutch online purchases in 2025 (Airwallex). b) 70% of shoppers prioritize payment methods when deciding where to shop (PYMNTS, 2025). c) A payment provider per country means a contract, a settlement schedule, and a reporting format per country. d) Marketplaces add another layer: every order has to split between the local vendors and the platform. The checkout decides whether a new market converts. Right? So how do you run one payment setup that follows the shopper across every market? Spree Commerce has everything you need and more: 1️⃣ Native Adyen integration: connect from the admin panel with two keys, in the free Community Edition. 2️⃣ Local methods picked per shopper location eg. iDEAL, Bancontact, SOFORT, digital wallets, 3D Secure. 3️⃣ One Adyen account: payments in over 150 currencies on one contract. A shopper in Amsterdam sees iDEAL first. A shopper in New York sees a card form and Apple Pay. No plugins. Not duct-taped add-ons. First-party features. Your Adyen contract stays yours. Spree Commerce adds no platform fee and no surcharge on any transaction. One checkout. Multiple markets. Local payment methods everywhere. Check out Spree Commerce and Adyen integration if this sounds interesting: spreecommerce.org/docs/integ… #ecommerce #opensource #payments #adyen #marketplace

Gartner has B2B buyers spending only 17% of their purchase time with sales reps. The other 83% is the storefront. And yet... A wholesale operator asked last week how to run per-customer pricing on one storefront without emailing price sheets every quarter. Why is per-account pricing still manual at most wholesale brands? a) The override logic lives in a spreadsheet a rep emails every quarter. b) Or there are multiple storefronts per segment, with parallel admin work. c) Or there is a pricing app bolted onto a SaaS built for one public price. The buyer should log in and see their contract price automatically. Right? Spree Commerce has everything you need in the free Community Edition: 1️⃣ Customer Groups: a tag on each customer record naming the segment. 2️⃣ Price Lists with Customer Group rules: the contract price for that segment. 3️⃣ Market rules: per-region wholesale tiers on the same segment. 4️⃣ Volume rules: quantity tiers above the segment price. A wholesaler with three tiers builds three Price Lists. Twelve segments? Not a problem. Twelve Price Lists. A distributor logged in from Germany sees the Germany distributor price. No plugins. Not duct-taped add-ons. First-party features. Need buyer organizations and approval workflows beyond pricing? The Enterprise Edition adds both on the same Customer Groups setup. For a DTC brand, adding a wholesale arm is a configuration change. Not a re-platforming project. One catalog. One checkout. Every account sees its own price. Check out Spree Commerce if this sounds interesting: spreecommerce.org/wholesale-… #ecommerce #opensource #b2b #wholesale #headless

Selling across EUR countries? €109 in Germany. €99 in France. €89 in Spain. Same product. Same currency. Different country. A brand selling across the EU asked last week how to do this on one storefront. Without multiple stores and admin panels. Cross-border pricing in EUR rarely means one price fits every country. Why? a) Shipping absorption. "Free shipping" in Germany costs more than in Spain. The same product quietly carries the price difference. b) VAT-inclusive display rounding. Germany 19%, France 20%, Spain 21%, Italy 22%. c) Local competitive pricing. Markets and direct rivals set the benchmark per country. Brands match locally rather than impose one global figure. d) Pricing power. Eurostat tracks consumer price levels across 36 European countries every year. Denmark sits 43% above the EU average. Bulgaria sits 40% below. The right number for the same SKU across markets is rarely identical. Right? So how do you charge different EUR prices in different countries without running a separate store per country? Spree Commerce has everything you need and more: 1) Markets: regions with their own currency, languages, payment methods, shipping rules. 2) Price Lists: an override layer that swaps the base price for shoppers in a given market. 3) Wholesale: stack a Market rule with a Customer Group rule for regional wholesale tiers. 4) Volume rules: per-market quantity tiers on the same Price List. A wholesale buyer logged in from Germany sees the Germany wholesale price. The same retail buyer from France sees the France retail catalog and price. No plugins. Not duct-taped add-ons. First-party features. For brands already selling in EUR across multiple eurozone markets, adding per-country pricing is a configuration change, not a re-platforming project. One catalog. Multiple countries. Multiple customer segments. Check it out: spreecommerce.org/cross-bord… #ecommerce #opensource #headless #b2b #crossborder

DoorDash, Ticketmaster, Faire, StockX, Zillow, Depop. Marketplaces you know. All run Next.js storefronts. Now you can have a stack like that for free. We documented what each of them is doing extremely well in a blog post: spreecommerce.org/nextjs-mar… And so can you with Spree Commerce open source natively supporting multi-vendor marketplace builds with a Next.js storefront on top. Each one of those platforms took millions of dollars and months, if not years, to build. Custom-built frontends. Custom integrations with the commerce backend. Custom payment-splitting logic. Engineering teams large enough to keep all of it shipping. Maisonette is a well documented case. They run a curated childrenswear marketplace with 65,000 products from hundreds of brands and raised $15 million in 2019. The 2018 build required Mirakl (six-figure annual subscription plus a GMV cut) for vendor orchestration, a custom Next.js storefront, a custom payment-splitting integration, and the engineers to wire it all together. Months of work. Seven figures before the first product page went live. That was 2018. In 2026, you can have what they all have on Spree Commerce, for a fraction of the cost, in weeks instead of months or years. Multi-vendor marketplace module: vendor onboarding, dashboards, commission management. Stripe Connect and Adyen for Platforms: automated split payments, escrow, payouts. Open-source Next.js eCommerce storefront on GitHub: Next.js 16, React 19, TypeScript 5, free. One-command install scaffolds backend and storefront in minutes. Headless REST API with OpenAPI 3.0: any frontend, any device, any agent. Fork the storefront. Connect your payment processor. Go live with a working marketplace in days. Not months. Not a year. The Next.js storefront experience that took DoorDash, Faire, and StockX millions to build is now a free starting point for everyone who comes next. That is the part that deserves attention. #ecommerce #opensource #marketplace #nextjs #headless

Interesting talk this week with @piotrkulpinski , the founder of @ossalternative, a community-driven directory of open source alternatives to SaaS tools. OpenAlternative traction tells a story: - 100,000 unique visitors in its first week - #1 on Hacker News - Crossed 1M unique visitors in 2025 That's a forward-looking signal. Self-hosted open source is having its moment. A few names you'll find on the site: 🤖 @n8n_io : AI-powered automation toolset. 🔍 @meilisearch : fast search engine for websites and apps 🌐 @hoppscotch_io : lightweight API client for testing endpoints on the fly 🤝 @twentycrm: open source CRM for contacts, deals, and collaboration 🛍️ @spreecommerce: headless eCommerce with REST API, TypeScript SDK, and Next.js storefront (integrated with Meilisearch, by the way) Why does this matter? Piotr said it better than I can, so I'll paraphrase him in three beats: 1️⃣ Transparency - it's open to humans and agents 2️⃣ Community - a pool of developers using and contributing 3️⃣ Freedom 👉 no vendor lock-in 👉 no dependence on one company's roadmap, 👉 full control over your own infrastructure and data That last one lands hardest for anyone who's ever been deplatformed, repriced, or regulated. Spree Commerce is built for exactly this moment. Open source and free to self-host anywhere. - Headless by default with a full REST API and OpenAPI 3.0 spec. - TypeScript SDK and a Next.js storefront ready to fork. - Bundled documentation in every Spree Commerce install. - AGENTS.md file maintained by the Spree Commerce team. Here's the quiet revolution part. AI coding agents love open source. Why? - They can read every line of source code. - Full documentation is local and indexable. - No API gates, no paywalls, no rate limits on what they can learn. - Years of community tutorials, GitHub issues, and public forums already live in their training. On a proprietary SaaS platform, the agent reasons from the outside in: outdated docs, blog posts, marketing pages, and whatever scraps the public API exposes. With Spree Commerce, the agent reads the codebase. Different level playing field. Whether you're building: - a B2B wholesale portal, - a multi-vendor marketplace, - a multi-tenant SaaS dealer network, - or a regulated commerce operation where the data has to stay yours ...self-hosted open source is no longer the compromise option. It's the strategic one. Open source is the new stack. Again. Try Spree Commerce for free. One command install: npx create-spree-app@latest my-store #ecommerce #opensource #headless #selfhosted #openalternative

If you're selling through Local Resellers, you must realize they are drowning in eCommerce busywork. Instead of selling your products. But there's a fix. This applies whether you sell through yoga studios, auto repair shops, local dealers, or franchisees. They're clicking around online stores all day. Managing inventory they don't understand. Shipping boxes. Handling returns. Chasing down orders. Meanwhile, what they actually want to do is teach, fix, sell and serve as many local customers as they can. You need to give them their time back. Streamline any eCommerce operations by centralizing them with a multi-tenant eCommerce platform. Shopify is great for single stores. But when you need to host hundreds of reseller storefronts - each with their own branding and customer base - while you maintain control, it’s pretty much impossible. So I just made a video breaking down exactly how to solve this with a multi-tenant eCommerce platform running on Spree Commerce. In this video, we cover: 👉 What multi-tenant ecommerce actually means   👉 How to allow resellers to focus on customers while you handle operations 👉 And how to get started without building everything from scratch GoDaddy runs 10,000+ stores on Spree Commerce. Your reseller network is probably smaller. But the platform is the same. The B2B businesses that win reseller ecosystems will be the ones that let a yoga studio owner keep teaching yoga. A car mechanic keep fixing cars. Keep a local dealer talking to as many customers as they can. So consider empowering hundreds of your resellers with their own, easy-to-use, hands-off online store. Host thousands of stores from a single multi-tenant eCommerce platform running on Spree Commerce. Manage them all from a single admin dashboard. This is the biggest growth hack a dealership business could unlock. Learn more about multi-tenant eCommerce here: spreecommerce.org/multi-tena… #ecommerce #opensource #multitenant #whitelabel #saas

API + MCP + CLI combo is the new UI. Everything is #headless now. Everything is just infra for AI. Marc Benioff announced Salesforce Headless 360 this week: all services exposed as APIs, MCP, and CLI. No browser required. @grinich said it bluntly: "death of UI." For 70 years we designed computer interfaces for humans. Now, the primary user of most enterprise software is about to be an agent. But who cares? UI was means to an end, anyway. Turns out some people care a lot. Software investors mostly. When every platform exposes itself as an API, MCP, and CLI, feature depth stops being a moat. A small team with a sharp prompt and the right scaffolding can replicate a lot of enterprise software in weeks. Billions of dollars of R&D, sitting in an API surface that any agent can read and any competitor can clone. You get agentic reach. You lose UI as a defensibility layer. Right? Not in eCommerce, though. OpenAI killed Instant Checkout. Stripe, Shopify, and PayPal all pulled support. Why? eCommerce is "complicated" - solving complex, mission-critical business problems across many domains? eCommerce (and business in general) is not a read-only query. It is money flows, inventory, tax, fulfillment, refunds, fraud, and regulatory edge cases across dozens of countries. Agents still need trusted commerce infrastructure underneath. Now more than ever. Which is where headless commerce gets interesting again. Spree Commerce has been building toward this moment. - Open source - full agent visibility. - Headless by default. - A REST API with OpenAPI 3.0 that passes all six agentic commerce protocols. - Full product docs bundled locally with every install. - Dedicated MCP server for the docs. - CLI for humans and agents. - AGENTS.md coding rules included. The platforms that win the next five years will not compete on admin UI. They will compete on how cleanly an agent can drive them. That is shipping today with Spree Commerce: github.com/spree/spree Open source is The Thing. Again. Try Spree for free with a single #CLI command: npx create-spree-app@latest my-store This takes you through an interactive setup: 1) Choose Full-stack (Backend + Next.js Storefront) or Backend only 2) Optionally load sample data (products, categories, images) Once complete, your store is running at http://localhost:3000/admin #ecommerce #opensource #AI #headlesscommerce #MCP

35% of the a16z Top 100 marketplaces run on Next.js storefronts. We've just open sourced a template you can use for free to go live within days. Fork it: github.com/spree/storefront Live: demo.spreecommerce.org/ It's a production-ready Next.js storefront. Not a demo. Not a "starter kit" you spend weeks customizing. A real storefront with everything you need to go live. - One-page checkout. - Guest and authenticated customer flows. - Coupons, gift cards, store credit. - Faceted product search powered by Meilisearch. - Multi-shipment support. - Native Stripe with Apple Pay and Google Pay. - GTM with GA4 eCommerce tracking. - Transactional emails. - SEO with structured data and OpenGraph. All of it ships when you fork the repo. Cross-border commerce works from day one. Country, currency, and language switching built into the URL structure via Spree Markets. No third-party localization plugins. No separate storefronts per region. Swap Stripe for Adyen or any other payment provider easily. The developer experience is where it gets interesting. Next.js 16, React 19, TypeScript, Tailwind CSS 4. The demo scores 90+ on Lighthouse / Google PageSpeed with perfect accessibility and SEO scores. Maisonette runs a curated childrenswear marketplace on this exact stack: Spree Commerce backend + Next.js storefront. Over 65,000 products, hundreds of vendors. What cost them $15M to build from scratch since 2018 is now a fork and a weekend of customization. That shift matters. - Install Spree Commerce with a single command. - Fork the storefront. - Get a working commerce application in minutes. For enterprise requirements like building a fully automated and scalable marketplace, let's talk. Try Spree for free: npx create-spree-app@latest my-store #ecommerce #opensource #nextjs #headlesscommerce #marketplace #typescript #react

The pivotal Spree Commerce 5.4 release is out: a 10x faster REST API, a TypeScript SDK, a Next.js eCommerce storefront, and built-in AI tooling: spreecommerce.org/announcing… Prototype today. Go live within days. Congrats to @damianlegawiec and the entire team. Some context to their achievement below. THE DEVELOPER STORY: A production-grade REST API with a TypeScript SDK and a Next.js eCommerce storefront ship as a single, deployable stack: ✅ TypeScript storefront teams fork the starter and ship within days. ✅ Mobile teams generate Swift or Kotlin clients from the OpenAPI spec. ✅ SaaS platforms call the REST API from whatever backend they already run. THE BUSINESS STORY: Launch cross-border eCommerce, multi-vendor marketplaces, and B2B commerce with a smaller team, shorter time to market, and a fraction of the upfront investment that custom builds demanded a few years ago. WHY IT MATTERS: 82% of developers evaluate new tools based on how quickly they can get a working prototype running. This latest Spree release lowers that barrier to near zero: one-command scaffolding, AI-native development with AGENTS.md, and multi-architecture Docker images. What can you build with Spree Commerce 5? ✅ Multi-store and cross-border commerce ✅ Multi-vendor marketplaces ✅ B2B wholesale portals ✅ Multi-tenant platforms Spree Commerce Enterprise Edition adds marketplace automations, advanced B2B, multi-tenancy, and enterprise security and support. Check out Spree Commerce 5.4: spreecommerce.org/announcing… Try Spree for free: npx create-spree-app@latest my-store #ecommerce #opensource #headlesscommerce #nextjs #b2bcommerce

6 agentic commerce protocols in the past 12 months have one pattern in common: ACP from OpenAI & Stripe, UCP from Google & Shopify, TAP from Visa, MCP from Anthropic, A2A + AP2 from Google. Every single one assumes a commerce platform exists underneath. They define how agents talk to commerce. They do not replace it. That's not a niche pattern. That's the foundation of the next economy being built in the open, right now. AI agents can recommend products, compare prices, and answer buying questions. But they cannot manage catalogs, orchestrate multi-gateway payments, calculate taxes across jurisdictions, or handle returns. Commerce is not a feature you bolt onto a chatbot. It's an infrastructure layer with nine interlocking capabilities that took decades to build. This is the Stripe parallel. Stripe defined how apps talk to payment infrastructure. These protocols are defining how agents talk to commerce infrastructure. The protocol layer creates massive value, but only because the infrastructure beneath it already exists. The platforms with complete REST APIs and machine-readable documentation are natively compatible with every protocol in this space. The ones relying on GraphQL-only interfaces or admin-panel-only workflows are architecturally excluded. Spree provides that layer: open-source, API-first, and purpose-built for the agentic commerce era. Complete REST API with OpenAPI 3.0 spec, AGENTS.md for AI context, MCP server, and every commerce operation accessible programmatically. Features get replicated. Infrastructure gets adopted. Here's the full breakdown on the Spree blog, including what AI agents actually need from commerce infrastructure and why OpenAI couldn't build it with $13 billion: spreecommerce.org/commerce-a… #ecommerce #opensource #ai #agenticcommerce #headlesscommerce #composablecommerce #api

#eCommerce Meta tried it. Google tried it 3 times. Twitter, Pinterest, Snapchat, TikTok. Now OpenAI tried it too. All failed. Why? Every big tech company in the last decade has attempted to own the the shopping experience and checkout. Every single one failed. Here's why. Meta — 5 years, billions invested, forced merchants onto native checkout, then abandoned it entirely in 2025. Google — Shopping Express, Buy on Google, Shopping app. Three attempts, three kills. Twitter, Pinterest, Snapchat — remember buy buttons? Pinterest had 10,000 merchants. Major retailers saw fewer than 10 purchases per day. TikTok Shop — spending billions and still fighting fulfillment disasters and seller backlash. Then OpenAI. No sales tax system. No fraud detection. 4% merchant fee stacked on top of existing processing costs. Out of millions of Shopify stores, about 12 integrated with ChatGPT's Instant Checkout before the feature was killed. Forrester's analyst said she was "shocked at the promises versus reality." The market's reaction was instant. The day the news broke, Expedia surged 12%. Booking Holdings gained 8%. Wall Street was saying: AI commerce disintermediation was overblown. The pattern is always the same. Whoever processes the payment becomes the merchant of record. That's the prize that drives repeat business and data ownership. But it also means owning refunds, fraud, tax compliance, customer service. Commerce is a full-time business. Big tech wanted the data and the business. But not the complexity. Commerce isn't a feature you bolt onto a chat window or a social feed. It's infrastructure. And infrastructure demands a dedicated commerce engine — like Spree. A deep dive into every failure — with the latest OpenAI fallout — in this blog post: spreecommerce.org/commerce-i… #opensource #marketplace #agenticcommerce

OpenAI Just Blinked: Nobody Seems to Want to Shop or Sell Inside ChatGPT. Here's why... Just 6 months after announcing OpenAI's vision of customers buying anything directly inside a chatbot, the company is quietly retreating — routing purchases through integrated retailers instead. Customers wouldn't buy - they were just browsing. Only ~12 (twelve) merchants out of millions on Shopify went live. OpenAI hadn't even built sales tax collection as of February 2026. We've built commerce & marketplace infrastructure at Spree for 15 years, and none of this surprises us. Here's the part the coverage is missing: Not every product can support an AI middleman. Chatbot or marketplace, the category math is the same. It's all mapped it out in the carousel: 1) AI takes all: insurance, SaaS, travel, cars. High margins, complex comparisons. Brokers and classifieds — your days are numbered. 2) AI discovers: gifts, fashion, beauty, home decor. Great discovery engine. But the transaction still lives on the commerce platform. 3) AI too late: groceries, staples, replenishment. Amazon and Walmart already own this. There's no discovery moment for AI to insert itself into. 4) AI too expensive: electronics, appliances, commodity hardware. A 4% AI tax on a $15 HDMI cable? The math doesn't work. The real question was never "Is AI shopping the future?" It was always "Which products can afford it?" OpenAI tried to own the full checkout — inventory, tax, fraud, payments, shipping, returns — and hit a wall on the basics. Commerce is infrastructure, not a chat feature. The next wave won't be chatbots replacing storefronts. It'll be AI-powered marketplaces built on flexible, open infrastructure — where the merchant controls the integration, the data, and the economics. If you're running the margin math on your own categories, we'd love to hear what you're seeing. Save this for your next platform strategy discussion. #ecommerce #opensource #agenticcommerce #marketplace #ai