> For the complete documentation index, see [llms.txt](https://docs.intelligems.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.intelligems.io/developer-resources/mcp-server/mcp-server-and-external-api-changelog.md). # MCP Server & External API Changelog A running log of what's new in the Intelligems MCP Server and External API. Newest updates are at the top. Unless marked (API), updates apply to both the MCP Server and the External API. *** ### November 2026 **New** * **Build any personalization through the MCP and API.** Everything you can build as a personalization in the app, your AI assistant can now build too: content edits, custom CSS/JS, redirects, every offer type (including free gift and Buy X Get Y), price changes, checkout blocks, post-purchase upsells, and onsite upsells. Paired with test creation, you can find an opportunity, launch a test, read the results, and roll the winner out as an always-on personalization without leaving your AI assistant. *(Nov 3)* * **Onsite upsells.** Create and edit onsite upsells (product recommendations on the product page or in the cart) as an A/B test or a personalization. Updates keep your existing upsells unless you send new ones. *(Nov 2)* * **Price-change personalizations.** Lower product prices for a specific audience with a personalization, using the existing price test tools to set prices. As in the app, price changes are decrease-only and can't be combined with an offer, checkout block, or onsite upsell. *(Nov 2)* * **Offer product lists.** Scope an offer to specific products instead of all products, for both tests and personalizations (amount off products, free gift, and Buy X Get Y), with variants included or excluded. A new `get_offer_products` tool reads the list back, and the External API adds `GET /v25-10-beta/experiences/{id}/offer-products`. *(Nov 3)* * **Personalization plays in Skills.** The built-in Skills now build personalizations directly, and `/roll-out` turns a winning test into an always-on personalization. *(Nov 3)* **Improved** * **Offer rules match the app.** A free gift offer now requires a gift tier, and a new volume discount requires at least two tiers. *(Nov 3)* **Fixed** * Offer, tier, and shipping discount IDs are now assigned automatically and are optional when creating offers, so offers on separate experiences can no longer collide. *(Nov 2)* *** ### October 2026 **New** * **Personalizations through the MCP.** Create and edit personalizations with content edits, custom CSS/JS, URL, theme, and template redirects, amount off order, volume discounts, free shipping, checkout blocks, and post-purchase upsells. *(Oct 29)* * **Checkout Blocks and Post-Purchase offers (API).** Manage checkout blocks and post-purchase offers with new `/v25-10-beta/experiences/{experienceId}/checkout-blocks` routes (GET, POST, PUT, DELETE), covering all six block types. *(Oct 28)* * **Create/Update Post-Purchase tests.** Create or edit Post-Purchase tests, including settings, discount, product selection, countdown, and styling. *(Oct 19)* * **Ready-to-use Skills, built into the MCP.** Common workflows are packaged directly into the Intelligems MCP, so your AI assistant can run them without any custom prompting. *(Oct 12)* **Improved** * **Plan and billing checks on test start.** Tests started through the MCP or API now go through the same plan and billing checks as the Start button in the app, with the same messaging. *(Oct 21)* * **Post-Purchase guardrails.** Each test group is limited to one post-purchase offer, the test's currency must match the store's, and the MCP now declines to change settings Shopify doesn't allow (like button text) instead of reporting them as updated. *(Oct 20)* * **Clearer ShopifyQL permission errors.** If your store hasn't granted ShopifyQL access, the MCP now tells you so directly instead of failing silently. *(Oct 18)* **Fixed** * Free Shipping offers can no longer be set up with an unsupported product-type configuration through the MCP. *(Oct 9)* * Shipping discounts can no longer be created with an unsupported dollar discount type, which previously produced invalid tests. *(Oct 7)* *** ### September 2026 **Improved** * **Upsell flag on checkout line items.** Checkout line-item data now identifies which items were upsells. *(Sep 1)* * **Split URL test validation.** The MCP now blocks invalid Split URL test configurations before they're created. *(Sep 21)* **Fixed** * Updating an experience no longer clears its custom-metric associations, and personalization audiences are now accepted correctly. *(Sep 24)* *** ### August 2026 **Milestone** * **Listed in the Claude Connectors Directory.** Intelligems is now available as an official connector in Anthropic's Claude directory. *(Aug 6)* **New** * **Create Checkout Block tests.** Build Checkout Block tests end to end through the MCP and API. *(Aug 27)* * **Create custom events.** Define new custom tracking events through the MCP. *(Aug 13)* * **Set a custom metric as the key metric.** Choose any custom metric as a test's primary success metric. *(Aug 18)* * **Currency targeting.** Target tests by currency through the MCP. *(Aug 3)* Currency targeting is now required when creating price tests, so tests launch with the right setup. *(Aug 26)* * **Activity log attribution.** Actions taken through the API or MCP now appear in your activity log, clearly labeled. *(Aug 10)* **Fixed** * Resolved a page targeting issue on tests created through the MCP. *(Aug 21)* *** ### July 2026 **New** * **Create price tests.** Set up price tests directly through the MCP and API. *(Jul 23)* * **Update price tests.** Edit existing price tests, including price modification fields. *(Jul 7)* * **Update Checkout Block tests.** Edit existing Checkout Block tests. *(Jul 31)* **Improved** * **Variation names and control labels in analytics.** Results now show variation names and clearly label the control group. *(Jul 9)* * **Smarter analytics tools.** Tool descriptions now include field-level metadata and synonyms, so AI assistants pick the right metrics more reliably. *(Jul 9)* * **Clearer organization data.** The `order_count` field in `get_organization` has been clarified. *(Jul 28)* **Fixed** * Active subscriptions were incorrectly returned as cancelled. This is now fixed. *(Jul 20)* * Fixed an invalid-parameters error in `search_products`. *(Jul 27)* * Fixed two bugs in the sitewide analytics endpoints. *(Jul 13)* *** ### June 2026 **New** * **Post-test analytics.** Pull the full post-test analytics view for completed tests. *(Jun 25)* * **Sitewide conversion funnel.** Access your store's conversion funnel data. *(Jun 4)* * **Stability-based test verdicts.** A new `test_status` field tells you whether a test's results have stabilized. *(Jun 30)* * **Custom event filtering.** Filter product-level test results by custom events. *(Jun 30)* **Fixed** * Profit-per-visitor metrics no longer silently fall back to revenue when COGS isn't configured. *(Jun 11)* *** ### May 2026 **New** * **ShopifyQL tools.** Query your Shopify store data directly through the MCP. *(May 9)* * **More sitewide analytics tools.** Additional sitewide tools added to the MCP. *(May 15)* * **BFCM tracking & benchmarking.** Access seasonal benchmarking data for BFCM and other major shopping events. *(May 18)* **Improved** * Clearer definitions for Probability to Beat Baseline (P2BB) and Probability to Beat Control (P2BC) in experiment analysis. *(May 28)* *** ### April 2026 **New** * **Order and line-item exports.** Export order-level and test line-item data. *(Apr 14)* * **Time series data (API).** Pull metrics over time. *(Apr 8)* * **Test history (API).** Retrieve your store's test history. *(Apr 27)* **Fixed** * Several analytics fixes, including errors when using start/end times, a 400 error when requesting AOV, mismatched conversion funnel filter results, and metric mix-ups. *(Apr 7)* *** ### March 2026 **New** * **Self-serve API keys.** Create your own API keys in the Intelligems app. *(Mar 10)* * **MCP context prompt.** A built-in context prompt helps AI assistants understand Intelligems data out of the box. *(Mar 10)* * **Custom metrics.** Custom metrics are now available. *(Mar 10)* * **Multiple filters and date range filtering.** Combine filters and set custom date ranges. *(Mar 13)* * **Sitewide analytics.** Access Performance, Audience Dive, and Order Distribution data. *(Mar 27)* * **POST experience data endpoint (API).** *(Mar 26)* **Improved** * The control variant now always appears first in API responses. *(Mar 2)* *** ### February 2026 **New** * **Write endpoints.** Take actions in your account, not just read data. *(Feb 12)* * **Sitewide analytics.** First sitewide metrics available. *(Feb 12)* * **Chart images.** Generate chart images of your results. *(Feb 23)* * **Estimated Monthly Revenue Increase.** Now available as a metric. *(Feb 27)* *** ### January 2026 **New** * **API key management.** Manage and revoke your API keys. *(Jan 9, Jan 20)* * **Rate limiting.** The External API now has rate limits to keep performance consistent for everyone. *(Jan 26)* **Improved** * Removed legacy fields that no longer returned reliable data. *(Jan 16)* *** ### December 2025 **Launch** * **External API (beta) and MCP Server launched.** Connect Intelligems to AI assistants like Claude, ChatGPT, and Gemini, or build your own integrations with the External API. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.intelligems.io/developer-resources/mcp-server/mcp-server-and-external-api-changelog.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.