ACP vs UCP: the two protocols that define how AI agents discover, evaluate, and purchase products from ecommerce stores. In the ACP vs UCP landscape, ACP, the Agentic Commerce Protocol, was developed by OpenAI and Stripe. UCP, the Universal Commerce Protocol, was developed by Google and Shopify. Both launched in early 2026 and both aim to standardize the end-to-end flow from product discovery to completed purchase by an AI agent. For ecommerce brands, the practical question in the ACP vs UCP decision is not which one to choose. It is how to support both.
This post covers how ACP and UCP work, how they differ in architecture and cost, what the March 2026 strategic shift in ACP means for Shopify merchants, and the implementation approach that positions your brand across both protocol ecosystems.
Is your Shopify store ready for AI agent purchases?
AI Advantage Agency helps ecommerce brands build the product data infrastructure, schema, and agentic commerce readiness needed to show up when AI agents shop on behalf of customers.
The Quick Take: ACP vs UCP at a Glance
| ACP (OpenAI + Stripe) | UCP (Google + Shopify) |
|---|---|
| Architecture: Centralized, platform-mediated | Architecture: Decentralized, merchant-hosted |
| Primary surface: ChatGPT ecosystem | Primary surface: Google AI Mode, Gemini, any agent |
| Cost: ~7.2% (4% platform fee plus processor) | Cost: ~3.2% (processor fees only) |
| Integration time: 2 to 4 hours | Integration time: 8 to 16 hours |
| Payment processor: Stripe only | Payment processor: Google Pay, PayPal (any processor planned) |
| Current focus: Product discovery (checkout scaled back March 2026) | Current focus: Full discovery and checkout, Cart and Catalog APIs added March 2026 |
💡 Pro Tip: The mental model that clarifies the distinction quickly: ACP is an agent-native checkout rail optimized for buying inside ChatGPT. UCP is an agent-native commerce language optimized for shopping across any agent surface. ACP and UCP are not competing for the same territory. They capture different demand. Build for both.
The Takeaway: ACP and UCP represent different philosophies for how AI agents connect to merchants. Both matter for ecommerce brands. The question is not which one to support but in what order and at what cost.
Table of Contents
→ What Is UCP and How Does It Work?
→ What Is ACP and How Does It Work?
→ The Core Design Philosophy Difference
→ Cost and Integration: What Shopify Brands Actually Pay
→ The March 2026 ACP Pivot: What Changed and Why It Matters
→ How Shopify Brands Should Implement ACP and UCP
→ The Bottom Line on ACP vs UCP
→ FAQ: Common Questions
What Is UCP and How Does It Work?
UCP is a decentralized, open commerce protocol that lets AI agents discover and purchase products directly from merchant servers without requiring registration with any central platform. Google and Shopify announced UCP at the National Retail Federation conference in January 2026. The launch coalition included over 20 companies: Shopify, Visa, Mastercard, Stripe, Wayfair, Target, Walmart, Best Buy, Etsy, American Express, Adyen, and others. (Google Blog, Agentic Commerce and UCP Announcement, January 2026.)
The defining technical characteristic of UCP is where merchant data lives. Merchants host a JSON profile at /.well-known/ucp on their own domain, following the same logic as robots.txt or sitemap.xml. Any AI agent can access that product information without prior registration or platform approval. This means UCP-enabled stores are discoverable by Google AI Mode, Gemini, and any other agent that implements the UCP specification, not just Google’s own surfaces.
UCP’s March 2026 update significantly expanded its scope beyond the initial launch. The update added a Cart API that lets agents add multiple items simultaneously and convert the cart to a checkout session, and a Catalog API that lets agents query live inventory, variants, and pricing directly from the merchant’s server rather than relying on pre-built product feeds. Payments process through AP2, the Agent Payments Protocol, using a mandate model based on Google Pay. For a full technical breakdown of UCP for Shopify specifically, see UCP for Shopify.
💡 Pro Tip: UCP onboarding for Shopify merchants is available through Merchant Center with six official platform partners: Shopify, Salesforce, Stripe, BigCommerce, Adobe, and Commerce Inc. If your store runs on Shopify, the integration path is straightforward through existing Merchant Center infrastructure. Start there rather than building a custom implementation from the UCP specification directly.
What Is ACP and How Does It Work?
ACP is a centralized commerce protocol where merchants submit product catalogs to OpenAI’s index, and AI agents discover and purchase products through the ChatGPT ecosystem. OpenAI and Stripe announced ACP in February 2026 with a vision of completing purchases seamlessly inside ChatGPT conversations. (OpenAI, Buy It in ChatGPT Announcement, February 2026.) Payments process exclusively through Stripe’s Shared Payment Token infrastructure.
ACP’s technical architecture uses four core API endpoints: Create Checkout for cart generation, Update Checkout for quantity and fulfillment changes, Complete Checkout for payment processing, and Cancel Checkout for transaction cancellation. The protocol uses REST HTTP only, secured through Bearer Token authentication and HMAC webhook signatures. ACP’s simpler technical surface is its primary integration advantage. Basic implementation completes in 2 to 4 hours compared to 8 to 16 hours for UCP.
The trade-off for that simplicity is platform dependency and cost. Merchants who implement ACP submit their catalog to OpenAI’s controlled index, where OpenAI manages ranking and discovery. Payments must route through Stripe. The combined cost of OpenAI’s 4% platform fee on top of Stripe’s processing fees brings the total to approximately 7.2% of transaction value. For brands doing significant volume, that fee differential against UCP’s 3.2% is material. The ACP specification is published under the Apache 2.0 license on GitHub and uses date-based versioning. (OpenAI, ACP GitHub Repository, 2026.)
The Core Design Philosophy Difference
UCP asks “can this be bought?” and starts with the merchant. ACP asks “should this be bought?” and starts with the agent. This distinction, articulated by Checkout.com’s analysis of the two protocols, captures the fundamental architectural divide between them. UCP is built on the premise that merchants should control their own data and be discoverable by any agent. ACP is built on the premise that a platform with massive user intent data, in this case ChatGPT, should mediate the discovery and recommendation layer.
UCP follows the open web model. Any agent that implements the UCP specification can query any UCP-enabled merchant. There is no gatekeeper, no ranking algorithm controlled by the protocol owner, and no submission process. Merchants who implement UCP retain full control of their data, their customer relationships, and their checkout experience. The protocol surfaces the capability; the agent decides whether to use it.
ACP follows the platform model. OpenAI controls which merchants appear in ChatGPT recommendations and in what order. A merchant cannot directly influence their ranking in ACP’s discovery layer the way they can optimize their UCP profile. For a broader look at how these protocols fit within the agentic commerce landscape, see what is agentic commerce.
| Dimension | UCP vs ACP |
|---|---|
| Discovery model | UCP: open, any agent. ACP: curated, platform controls ranking. |
| Merchant control | UCP: merchant owns data and customer relationship. ACP: platform mediates both. |
| Transport | UCP: REST, JSON-RPC, MCP, A2A, AP2. ACP: REST HTTP only. |
| Analogy | UCP is to agentic commerce what HTTP is to the web. ACP is to agentic commerce what the App Store is to mobile apps. |
💡 Pro Tip: The open vs. platform distinction has long-term strategic implications beyond current traffic. UCP-enabled merchants are not dependent on any single platform’s algorithm for AI agent discovery. ACP-enabled merchants have access to ChatGPT’s user base but accept platform dependency. Both considerations matter. Build for UCP first if merchant control and cost efficiency are priorities. Add ACP for ChatGPT distribution reach.
Cost and Integration: What Shopify Brands Actually Pay
The cost differential between ACP and UCP is significant at scale and worth calculating against your actual revenue before committing to an implementation sequence. UCP carries approximately 3.2% in processor fees with no platform fee. ACP carries approximately 7.2% total, comprising OpenAI’s 4% platform fee plus Stripe’s processing fees. (Stellagent, UCP vs ACP Comparison, April 2026.)
At $100,000 in monthly agent-driven revenue, the cost difference is $4,000 per month. At $500,000 in monthly agent-driven revenue, it is $20,000 per month. These numbers are projections on hypothetical agent-driven volume, not guaranteed outcomes. Agent-driven commerce is early-stage for most SMB brands, so the absolute dollar impact is modest now. The reason to understand it now is that protocol choices made at low volume become structural decisions at high volume.
Integration complexity favors ACP for teams without dedicated engineering resources. ACP’s 2 to 4 hour integration time against UCP’s 8 to 16 hours is a real difference for smaller brands. However, Shopify’s native UCP support through Merchant Center compresses the UCP integration significantly for Shopify merchants specifically. If your store is on Shopify, the UCP integration path is substantially easier than the raw specification time estimate suggests.
The March 2026 ACP Pivot: What Changed and Why It Matters
In March 2026, OpenAI scaled back ACP’s Instant Checkout feature to focus on product discovery instead. The trigger was adoption: out of Shopify’s millions of stores, only approximately 12 merchants had activated ACP checkout. OpenAI acknowledged that the initial version of Instant Checkout did not offer the level of flexibility merchants needed. (Digital Commerce 360, March 2026.)
This pivot has two practical implications for ecommerce brands evaluating ACP vs UCP. First, ACP’s current primary value for most merchants is product discovery within ChatGPT, not checkout completion. Implementing ACP today primarily means getting your products surfaced in ChatGPT product recommendations, not enabling AI agent purchases through ChatGPT. Second, the checkout gap between UCP and ACP has widened. UCP’s March 2026 update added Cart and Catalog APIs that expanded its checkout capabilities while ACP stepped back from that space.
OpenAI’s expanded discovery partnerships with Target, Sephora, Nordstrom, Lowe’s, Best Buy, Home Depot, and Wayfair signal that ACP’s near-term roadmap prioritizes large retailer integration over SMB checkout infrastructure. For SMB Shopify brands, UCP currently offers the more complete end-to-end agentic commerce solution. ACP remains valuable for ChatGPT discovery reach but should not be the primary implementation focus if checkout capability is the goal. For how agentic commerce readiness maps to your full stack, see the agentic commerce readiness checklist.
💡 Pro Tip: Monitor OpenAI’s ACP roadmap updates directly. The protocol is evolving quickly and the checkout pivot may reverse as adoption grows. The Apache 2.0 specification on GitHub is updated with date-based versioning. Check the changelog before making implementation decisions rather than relying on coverage from earlier in 2026 that may not reflect the current state.
How Shopify Brands Should Implement ACP and UCP
The standard recommendation across the agentic commerce ecosystem is a dual-protocol strategy: build a shared product data backend, then implement both UCP and ACP endpoints on top of it. This approach avoids maintaining two separate data pipelines and ensures that product catalog updates, inventory changes, and pricing adjustments propagate to both protocol surfaces simultaneously. (Stellagent, April 2026.)
For Shopify brands, the ACP vs UCP implementation sequence that makes the most sense in mid-2026 is UCP first, ACP second. UCP offers native Shopify support through Merchant Center, the more complete checkout capability, lower transaction costs, and open discovery across any agent surface. ACP adds ChatGPT distribution reach and is worth implementing once UCP is live and functioning. Do not delay UCP implementation waiting for ACP to mature further.
Both protocols require the same underlying product data foundation: accurate, machine-readable product data with complete schema markup, real-time inventory signals, and consistent pricing between your on-page product data and any feed or profile you expose to agents. An agent that discovers a product priced at $49.99 on your UCP profile but finds $59.99 on the product page will flag the data as unreliable and exclude the product from its recommendations. For the product data requirements that underpin both protocols, see attribute-rich product data for ecommerce.
The Bottom Line on ACP vs UCP
ACP and UCP are complementary protocols that together define how AI agents will discover and purchase products across the two largest AI commerce surfaces: Google and ChatGPT. The ACP vs UCP decision is not a choice between competing protocols. Neither is optional for ecommerce brands that want full coverage as agentic commerce scales. The question is sequencing, not choice.
Implement UCP first. It offers the more complete end-to-end agentic commerce capability in mid-2026, lower transaction costs, native Shopify support, and open discovery across any agent that implements the specification. Add ACP for ChatGPT distribution reach once your UCP implementation is stable. Both require the same product data foundation, so the shared backend investment applies to both.
The brands that build agentic commerce infrastructure now are establishing the data quality, schema completeness, and protocol readiness that will determine their visibility when AI agent shopping scales to mainstream volume. The window to build that foundation before the channel becomes crowded is open now. The protocols exist. The implementation paths are documented. The cost of waiting is compounding invisibility in the channel that will reshape ecommerce discovery over the next three years.
🎯 Ready to Make Your Shopify Store Visible to AI Shopping Agents?
AI Advantage Agency helps ecommerce brands build the product data infrastructure, schema markup, and protocol readiness needed to show up when AI agents shop on behalf of customers.
→ Book a Free Agentic Commerce Audit
We review your current product data setup and show you exactly what to implement first.
Frequently Asked Questions About ACP vs UCP
What is the difference between ACP and UCP?
ACP (Agentic Commerce Protocol) is a centralized protocol developed by OpenAI and Stripe where merchants submit catalogs to OpenAI’s platform and products are discovered through ChatGPT. UCP (Universal Commerce Protocol) is a decentralized protocol developed by Google and Shopify where merchants host product data on their own servers and any AI agent can discover it. ACP is platform-mediated; UCP is open and merchant-controlled.
Which is better for Shopify brands, ACP or UCP?
UCP is the better starting point for Shopify brands in mid-2026. It offers native Shopify support through Merchant Center, lower transaction costs (approximately 3.2% vs 7.2% for ACP), a more complete end-to-end checkout capability, and open discovery across any agent surface. ACP adds ChatGPT distribution reach and should be implemented second, once UCP is live.
How much does it cost to implement ACP vs UCP?
UCP costs approximately 3.2% of transaction value in processor fees with no platform fee. ACP costs approximately 7.2% total, comprising OpenAI’s 4% platform fee plus Stripe’s processing fees. At significant agent-driven revenue volume, this difference is substantial. Integration time also differs: UCP requires 8 to 16 hours, ACP requires 2 to 4 hours.
What happened to ACP’s checkout feature in March 2026?
In March 2026, OpenAI scaled back ACP’s Instant Checkout feature to focus on product discovery instead. The reason was low adoption: approximately 12 of Shopify’s millions of merchants had activated checkout. OpenAI acknowledged that the initial version lacked the flexibility merchants needed. ACP’s current primary value for most merchants is product discovery within ChatGPT rather than checkout completion.
Can a Shopify store implement both ACP and UCP?
Yes, and dual-protocol implementation is the recommended approach. The standard strategy is to build a shared product data backend with real-time inventory and pricing, then implement both UCP and ACP endpoints on top of it. This avoids maintaining two separate data pipelines and ensures catalog updates reach both protocol surfaces simultaneously.
When were ACP and UCP launched?
UCP was announced by Google and Shopify at the National Retail Federation conference in January 2026. ACP was announced by OpenAI and Stripe in February 2026. Both protocols have been updated since launch, with UCP adding Cart and Catalog APIs in March 2026 and ACP pivoting away from checkout toward product discovery in the same month.
What product data do I need before implementing ACP or UCP?
Both protocols require accurate, machine-readable product data with complete schema markup, real-time inventory signals, and consistent pricing between your on-page product data and any feed or protocol profile. An agent that finds mismatched pricing between your UCP profile and product page will flag your data as unreliable and exclude your products from recommendations.
Does UCP work with payment processors other than Google Pay?
UCP currently supports Google Pay with PayPal planned. The protocol is designed to support any payment processor over time, which is part of its open architecture philosophy. ACP requires Stripe exclusively, as Stripe co-developed the protocol and its Shared Payment Token infrastructure is the payment layer for all ACP transactions.
Is ACP or UCP required for appearing in AI shopping results?
Not yet required, but increasingly important for AI agent-driven discovery and purchase. UCP is required to participate in Google AI Mode’s agentic shopping experience. ACP is required to participate in ChatGPT’s product discovery ecosystem. Brands without either protocol are invisible to AI agents that use these standards to discover and transact with merchants.
How is UCP different from MCP?
MCP (Model Context Protocol) is Anthropic’s protocol for connecting AI agents to external data sources and tools. UCP is Google and Shopify’s protocol for standardizing how AI agents discover and purchase products from ecommerce stores. MCP handles the broader agent-to-data connection layer. UCP handles the specific commerce transaction layer. UCP supports MCP as one of its transport methods.
