How is this different from connecting Gorgias MCP directly?

Connected directly, an agent sees every tool the server exposes and holds the credential itself. Through TrustGate you choose which tools each agent gets, and every call is logged against the agent that made it.

What if the Gorgias MCP API changes?

We track the Gorgias MCP schema, so the tool catalogue stays current. Tools you have already allowed keep working, and anything new stays denied until you enable it.

Can I give different agents different Gorgias tools?

Yes. Access is set per agent: one can get a handful of read tools, another the full set. You grant the whole MCP or single tools within it.

Where can I see what my agents did in Gorgias?

Every call is logged against the agent that made it, with the tool it used. Activity can be sent on to your own monitoring.

Do I need to run the Gorgias MCP server myself?

No. TrustGate connects to the Gorgias MCP server and sits in front of it. Your agent only talks to your TrustGate URL, never to Gorgias directly.

Connect Gorgias MCP to your AI agents, securely.

Q: Is Gorgias MCP secure?

The server itself exposes 165 tools, and some of them can write to your Gorgias account. Behind TrustGate an agent only sees the tools you allow, requests and responses are inspected, and every call is logged. Anything outside the allowlist is denied before it reaches Gorgias.

One secured endpoint for every Gorgias call your AI agents make, scoped per user, inspected, and logged.

Get started for free

Get a demo

Try Gorgias MCP now

What your agents can ask once Gorgias MCP is connected.

“Find the ticket where the customer mentioned a defective hinge”
“Show me this ticket with the full conversation and the customer record”
“What's our automation rate this month, broken down by channel?”

“Draft a help center article about our return policy and leave it as a draft”
“Which guidance articles are enabled for the AI Agent right now?”
“Preview how the AI Agent would answer 'Where is my order?'”

Select the tools your AI agents can call

165 tools from the Gorgias MCP schema.

Showing 165 tools

get_ticketreadGets a single ticket with tags, summary and, by default, its messages embedded, so one call gives the full conversation. Can inline the full customer record.
list_ticketsreadLists tickets by metadata filters: customer, external ID, rule, trashed state, or a Gorgias view that pre-filters by status, channel and assignee.
search_ticketsreadFull-text Elasticsearch search across ticket subjects, messages and customer fields, with highlighted match snippets and an optional view-DSL filter.
add_internal_notewritePosts an internal note on a ticket, not visible to the customer, optionally @mentioning teammates who then receive a notification.
snooze_ticketwriteSnoozes a ticket until a given datetime so it leaves default inbox views, optionally posting the reason as an internal note.
escalate_ticketwriteEscalates a ticket in one call: assigns it to a team, adds the escalated tag and leaves an internal note recording who escalated, when and why.
create_messagehigh riskPosts a message on a ticket and, by default, delivers it to the customer. Supports forwarding to another address and recording without sending.
reply_and_closehigh riskPosts a customer-facing reply and closes the ticket in one call, the Send and Close pattern.
create_tickethigh riskOpens a new ticket with an initial message, auto-creating the customer if unknown. Outbound email is recorded without sending unless send is set to true.
update_ticket_custom_fieldshigh riskReplaces a Gorgias ticket's custom fields.
update_ticketshigh riskUpdates one or many Gorgias tickets at once, via an async bulk job.
get_jobreadChecks the status of an async Gorgias job from apply_macro or update_tickets.
list_ticket_audit_eventsreadLists the safely summarized audit trail for one ticket.
get_help_center_articlereadFetches a help center article with its raw HTML, publication status and storefront visibility, in published or draft version.
list_help_center_articlesreadLists the articles in a help center with their publication status and visibility, public or unlisted.
list_help_center_categoriesreadLists the categories, or sections, of a help center.
list_help_centersreadLists the account's public FAQ help centers with their domain, locales and storefront status. Some accounts run several, one per brand or domain.
create_help_center_articlewriteCreates a help center article, as a draft by default so a person can review it before it goes live.
update_help_center_articlewriteSaves a new draft version of an existing article. The live storefront version is unchanged until it is published.
publish_help_center_articlehigh riskMakes the draft the live version, published and listed on the public storefront.
discard_help_center_article_drafthigh riskDiscards pending draft edits. For a published article the live version is restored; for an article that was never published, discarding its only draft removes it entirely.
unlist_help_center_articlehigh riskHides an article from the storefront without deleting it.
get_help_center_themereadFetches a help center's theme and storefront settings.
update_help_center_themehigh riskUpdates a help center's theme and storefront settings.
get_help_center_extra_htmlreadFetches a help center's custom HTML blocks for a locale.
update_help_center_extra_htmlhigh riskUpdates a help center's custom HTML blocks for a locale.
get_guidancereadFetches a guidance article, whose translation block carries two independent axes: publication status, draft or published, and AI agent status, enabled or disabled.
list_guidancesreadLists guidance articles in a store's knowledge hub with their publication and AI agent status.
list_guidance_templatesreadLists the curated best-practice guidance templates used for initial AI Agent setup or as a coverage check against an existing guidance base.
get_skillreadFetches a single skill with its full content and linked intents, in the published or latest draft version.
list_skillsreadLists the skills for a store, meaning articles linked to intents, with their visibility status and full intent list.
list_intentsreadLists the available intents for a store with whether each is linked to a published skill, unlinked, or set to hand over to a human agent.
create_draft_guidancewriteCreates a guidance article as a draft. It is saved but not published and not yet enabled for the AI Agent.
update_draft_guidancewriteSaves a new draft version of an existing guidance. Its current AI Agent availability carries forward unchanged, so saving never enables or disables it.
disable_guidancewriteStops the AI Agent using a guidance by flipping its status to disabled. The article content is preserved.
create_draft_skillwriteCreates a skill as an unlisted draft with its linked intents.
update_draft_skillwriteSaves a new draft version of an existing skill, carrying the current visibility status forward. A passed intent list replaces the current set.
publish_guidancehigh riskPublishes a draft guidance, making it the live version and enabling it for the AI Agent to use with customers.
publish_skillhigh riskPublishes a draft skill, setting it current and public, and preserving the latest draft's intent list.
discard_draft_guidancehigh riskDiscards the pending draft on a guidance and restores the last published version. Any unpublished edits are lost.
list_skill_versionsreadLists retained versions of a skill translation, newest first.
get_skill_versionreadFetches one exact retained skill revision with its stored HTML content.
get_action_templatereadFetches a template's full recreatable definition, its steps, transitions, triggers, entrypoints and inputs, for inspection before recreating it with edits.
list_action_templatesreadLists the pre-built action templates with their category, instructions, required data, available apps and whether they overlap a built-in capability.
get_support_actionreadFetches a support action's full configuration plus its last three execution summaries for diagnostics.
list_support_actionsreadLists the AI agent support actions configured for a store, including built-in capabilities that cannot be edited or disabled.
create_action_from_templatewriteDeploys a pre-built action template to a store, disabled by default. It refuses templates that duplicate an always-on built-in capability.
create_support_actionwriteCreates an AI agent support action from explicit steps and transitions, with every entrypoint deactivated until someone enables it.
update_support_actionwriteMerges changes into an existing support action's configuration. Only top-level keys included in the change set are modified.
disable_support_actionwriteDeactivates a support action's entrypoints while preserving its configuration.
enable_support_actionhigh riskEnables a disabled support action so the AI agent can invoke it against real conversations.
convert_to_advanced_viewhigh riskConverts a support action to the Advanced View, unlocking the full step editor. This is one-way and irreversible: the action can never be downgraded.
list_support_action_executionsreadLists a support action's Workflows execution history, filtered and paged.
get_support_action_executionreadReads one persisted support-action execution and its HTTP audit evidence.
duplicate_support_actionwriteCreates an exact, disabled copy of an existing support action.
delete_support_actionhigh riskPermanently deletes a store-scoped support action.
get_agent_configurationreadGets the full AI Agent configuration for a store: help center IDs, tone of voice, channel settings, monitored integrations and other parameters.
list_agent_configurationsreadLists the AI Agent store configurations on the account with a summary per store.
preview_ai_agentreadSends a customer message through the AI Agent in test mode, returning its reply, outcome, reasoning and the knowledge sources it used, without sending real messages or running real actions.
preview_tone_of_voicereadPreviews an AI Agent reply with a candidate tone of voice, for iterating on wording before committing it.
update_tone_of_voicewriteSets the AI Agent's personality preset for a store, or custom tone instructions when the preset is Custom.
ai_agent_setup_completionwriteMarks the AI Agent setup wizard complete for a store, or reports the current wizard state.
enable_ai_agent_on_channelhigh riskTurns the AI Agent on for chat, email or SMS and replaces that channel's monitored-integration list, putting the AI in front of real customer conversations.
update_excluded_handover_topicshigh riskSets the topics the AI agent must not handle and should hand to a human instead. Passing an empty list clears every excluded topic.
preview_ai_agent_with_live_actionshigh riskTests the configured AI Agent through a shopper conversation with live actions enabled.
continue_ai_agent_previewwriteSends a follow-up customer turn to an existing mock-only test session.
continue_ai_agent_preview_with_live_actionshigh riskContinues a known-live AI Agent test conversation by submitting a shopper message.
get_ai_agent_preview_sessionreadInspects allowlisted status and conversation logs for a V3 test session.
get_ai_reasoningreadReads safely allowlisted AI Agent reasoning for a ticket or V3 preview.
list_ai_agent_preview_optionsreadLists available store, channel, and integration context for AI Agent test sessions.
compare_ai_agent_previewswriteCompares two to four mock-only AI Agent test conversations sequentially.
compare_ai_agent_previews_with_live_actionshigh riskCompares independent AI Agent test conversations with live actions enabled.
get_macroreadGets a single macro, including its full actions array, for inspection before applying or editing it.
list_macrosreadLists macros with a compact summary each: intent, language, usage counter, action count and archived flag, optionally ranked by relevance to a ticket.
create_macrowriteCreates a macro, active and not archived. Its action list can set response text, tags, status, assignee, subject, snooze, internal notes or send email.
update_macrowriteUpdates fields on an existing macro. A passed actions array replaces the existing list; partial action updates are not supported.
apply_macrohigh riskApplies a macro to a ticket through Gorgias's server-side endpoint, running whatever bundle of actions the macro carries, which can include sending email.
archive_macroshigh riskHides macros from active use without deleting them.
restore_macroshigh riskReturns archived macros to active use without changing their contents.
get_rulereadFetches a single automation rule with its editable fields, including the JavaScript source, for review before updating it.
list_rulesreadLists the account's helpdesk automation rules with a compact summary each: event types, priority and enabled flag.
create_rulewriteCreates an automation rule from JavaScript source and trigger events. It always lands disabled, which the server states as a hard invariant rather than an option.
disable_rulewriteSoft-disables an automation rule, preserving its configuration.
enable_rulehigh riskRe-enables a disabled automation rule so it fires on its configured ticket events again.
update_rulehigh riskUpdates an existing rule's fields, including the JavaScript body that runs automatically, its trigger events and its execution priority.
simulate_rule_eventreadSimulates a helpdesk rule firing on a ticket — nothing is dispatched.
list_rule_audit_eventsreadReturns the audit trail for one helpdesk rule: who created, updated or deleted it, and when.
get_reporting_statsreadFetches live operational stats from the Reporting API. Scope, measures and required filters must come from list_metric_cards and the schema, never from guesswork.
get_reporting_stats_schemareadReturns the available scopes, measures, dimensions and filters for the Reporting Stats API. The full schema covers more than 40 scopes, so narrowing by scope is recommended.
list_metric_cardsreadLists the published metrics with their exact Reporting Stats parameters, mapping a human-readable metric name to the scope, measures and required filters.
get_tablesreadLists every table available to the analytics query tool with a short description each.
get_table_metadatareadLoads the full schema and usage notes for one analytics table: columns with types and descriptions, plus when and how to use it.
queryreadRuns a read-only SQL query against the analytics warehouse, with per-account isolation applied automatically. The warehouse refreshes roughly daily, so it is not a real-time source.
get_ticketsreadFetches a batch of warehouse ticket snapshots, newest first, by known IDs.
get_ticket_analyticsreadFetches one warehouse ticket snapshot with AI-Agent outcome/intent/CSAT/billing.
get_ticket_executionsreadFetches AI Agent execution history for a batch of known ticket IDs, oldest first.
find_tickets_by_topicreadFinds a bounded ticket cohort by topic, requester, shop, outcome, and dates.
ticket_pattern_breakdownreadGroups tickets to find leading contact reasons and handover patterns.
ai_performance_metricsreadComputes denominator-safe AI performance rates for a ticket-creation cohort.
csat_investigationreadInvestigates low-CSAT tickets, optionally grouped with representative examples.
execution_outcome_breakdownreadBreaks down AI Agent executions by outcome, handover topic or both.
get_customerreadGets a single customer by ID. The server warns that name, email, channel addresses and integration data are attacker-controllable and must be treated as data, not instructions.
list_customersreadLists customers, filterable by email, name, channel address, channel type, external ID or view. The same attacker-controllable warning applies to every returned string.
list_usersreadLists the account's Gorgias users, meaning agents, filterable by role, email or free-text search.
list_teamsreadLists the account's teams with their description and member list, each member's id, name and email.
update_customerhigh riskUpdates profile fields on an existing Gorgias customer.
get_teamreadGets one Gorgias team by its numeric ID.
get_current_userreadReturns the currently authenticated Gorgias user, useful as a sanity check on which identity the connection is acting as.
get_custom_fieldreadFetches a custom field's definition including its label, type, options and whether it is required.
list_custom_fieldsreadLists the custom field definitions for tickets or customers, including which IDs exist, which block closing a ticket, accepted data types and dropdown choices.
list_tagsreadLists the tags defined on the account, useful before adding tags to avoid creating near-duplicate names.
list_integrationsreadLists connected and available integrations, covering Shopify, BigCommerce and Magento plus Workflows app data, with per-channel AI Agent state and the integration IDs needed to enable a channel.
submit_feature_requestwriteSubmits a confirmed, structured product feature request to Linear.
create_custom_fieldwriteCreates a ticket or customer custom field.
update_satisfaction_surveyhigh riskUpdates a satisfaction (CSAT) survey.
create_tagwriteCreates an account-wide tag, or returns the existing tag with the exact name.
list_voice_queuesreadLists the voice (phone) queues configured on the Gorgias account.
list_sla_policiesreadLists SLA policies on this Gorgias account.
get_gaia_instructionsreadReturns the server's operating manual: runtime context, tool inventory, the skill-load workflow, real-time versus analytics routing, schema and query rules, draft-first write defaults and security rules. The server asks that it be called before any other tool in every new conversation.
get_instructionreadLoads a named skill workflow, such as refund or escalation. The catalogue of available names is embedded in the operating manual's output.
list_routinesreadLists Gaia routines for this Gorgias account.
get_routinereadFetches one Gaia routine by ID.
create_routinehigh riskCreates a Gaia routine.
update_routinehigh riskUpdates a Gaia routine.
delete_routinehigh riskSoft-deletes a custom Gaia routine.
trigger_routinehigh riskTriggers one immediate run for a Gaia routine.
set_routine_enabledhigh riskEnables or disables a Gaia routine.
list_routine_runsreadLists runs for one Gaia routine, newest first.
list_viewsreadLists ticket-list views.
get_viewreadGets a ticket-list view by numeric ID.
create_viewwriteCreates a ticket-list view.
update_viewhigh riskPartially updates a ticket-list view; supplied fields replace their current values.
delete_viewhigh riskDeletes a ticket-list view.
list_view_sectionsreadLists view sections, ordered by creation time descending.
get_view_sectionreadGets a view section by numeric ID.
create_view_sectionwriteCreates a view section.
update_view_sectionhigh riskPartially updates a view section.
delete_view_sectionhigh riskDeletes a view section.
list_snippetsreadLists reusable AI Agent snippets for one Shopify store.
get_snippetreadFetches one reusable snippet and its visible content.
list_snippet_sourcesreadLists domain, URL, or file sources that produce snippets.
get_snippet_sourcereadFetches one snippet source and the allowlisted articles it produced.
set_snippet_statushigh riskEnables or disables a reusable snippet after fresh explicit human confirmation.
search_productsreadSearches Shopify products connected to this Gorgias account.
get_product_custom_inforeadFetches AI-agent custom info (merchant-authored context) for Shopify products.
list_product_metafield_definitionsreadLists Shopify metafield definitions configured on this store.
set_product_custom_infohigh riskSets or clears the AI Agent "Custom Product Information" for one Shopify product.
pin_metafield_definitionwritePins or unpins a Shopify metafield definition for the AI agent.
list_opportunitiesreadLists pending AI Agent Product Opportunities for a Shopify store.
get_opportunityreadGets one pending AI Agent Product Opportunity for a Shopify store.
get_app_detailsreadGets customer-safe details for one Gorgias App Store app.
list_http_integrationsreadLists custom HTTP integrations (Settings -> Integrations -> HTTP) for this account.
get_http_integrationreadFetches one custom HTTP integration by id.
list_http_integration_logsreadLists recent requests (the request log) for one custom HTTP integration.
list_widgetsreadLists the sidebar widgets configured on this Gorgias account.
get_widgetreadReads one sidebar widget, including its full render template.
update_widgethigh riskUpdates a sidebar widget; the template field replaces the whole template object.
create_widgetwriteCreates a new sidebar widget.
delete_widgethigh riskDeletes a sidebar widget (soft-delete upstream).
get_segmentation_conditionsreadReturns the catalog of audience fields, aggregates, operators and combinators for the account.
list_segmentsreadLists this account's saved audience segments (id, name, store, conditions).
get_segmentreadReturns one saved audience segment with its ID, name, store and full conditions.
get_saved_segment_sizereadReturns how many people a saved segment matches right now, as an aggregate count only.
get_direct_audience_sizereadCounts people matching ad-hoc segmentation conditions without saving a segment.
query_audience_membershipreadAnswers "who is in segment X?" with the segment's current total only.

What Gorgias MCP exposes

Gorgias MCP groups its 165 tools into 16 categories, each with a suggested starting policy.

Category	Covers	Access	Suggested policy
Guidance & Skills	Guidance articles in the knowledge hub and the skills that link articles to customer intents, with their templates, draft creation and editing, publishing, disabling, draft discarding and retained skill versions	Read / Write	Guidance and skills are what the AI Agent reads before answering a customer, so publishing one changes live customer-facing behaviour. Drafts are safe; publish_guidance also enables the article for the AI Agent
Tickets	Reading, searching and listing tickets with their messages and customer record, internal notes, tags, snoozing, escalation, single and bulk field updates, custom field values, the ticket audit trail, customer replies and new ticket creation	Read / Write	create_message and reply_and_close send real messages to customers by default; create_ticket records without sending unless send is set. update_tickets runs as an async, non-atomic bulk job, so check the result with get_job
Support Actions	AI agent support actions and the pre-built templates they can be deployed from, with creation, updating, enabling, disabling, duplication, deletion, execution history and conversion to the advanced step editor	Read / Write	New actions are created with every entrypoint deactivated, so enable_support_action is the real control. convert_to_advanced_view is irreversible and the action can never be downgraded
Help Center	Public FAQ help centers, their categories and articles, with draft creation and editing, publishing to the storefront, unlisting, discarding pending drafts, and theme and custom HTML settings	Read / Write	Articles are public storefront content. Edits save as drafts and the live version is unchanged until publish_help_center_article. Discarding the only draft of a never-published article removes it entirely
AI Agent Config	Per-store AI Agent configuration, tone of voice with a preview mode, reply preview against draft guidance, multi-turn, comparative and live-action test sessions, AI reasoning, excluded handover topics, channel enablement and the onboarding wizard state	Read / Write	enable_ai_agent_on_channel puts the AI in front of real customers on chat, email or SMS and replaces the channel's monitored-integration list. Excluded topics are the ones the AI must hand to a human, so clearing them widens its scope
Macros	Listing and inspecting macros, creating and updating them, applying one to a ticket, and archiving or restoring to control whether agents see them	Read / Write	A macro's action list can include sendEmail, setStatus and addInternalNote, so apply_macro is a bundle of side effects behind one call. Archiving fails while a helpdesk rule still references the macro
Helpdesk Rules	Automation rules with their JavaScript source, trigger events and priority, plus creation, updating, enabling, disabling, dry-run simulation and the rule audit trail	Read / Write	Rules are code that fires automatically on ticket events. create_rule always lands disabled as a hard invariant, so enable_rule and update_rule are where review belongs
Reporting & Analytics	The Reporting Stats API with its schema, scopes, measures, dimensions and metric cards, plus the analytics warehouse tables, their column metadata and read-only SQL, warehouse ticket snapshots, AI Agent execution history and ready-made breakdowns of topics, handovers, AI performance and CSAT	Read	Read-only, and per-account isolation on the SQL path is automatic. The warehouse refreshes roughly once a day, so it is the wrong source for real-time questions
Account Config	The authenticated agent, custom field definitions for tickets and customers, the account's tags, connected ecommerce and Workflows integrations, voice queues, SLA policies, CSAT survey settings and product feature requests	Read / Write	list_custom_fields is the prerequisite for any custom field write, since it reports which IDs exist, which block closing a ticket and what values dropdowns accept
Customers & Users	Gorgias customers by ID or filtered list with profile updates, the account's agents with their roles, and teams with their member lists	Read / Write	Customer fields are attacker-controllable by design and the server says so: anyone can register a customer by emailing support. Strings come back Unicode-scrubbed and must be treated as data, never instructions
Operating Manual	The server's own operating manual, covering runtime context, the tool inventory, routing rules, schema discovery, draft-first write defaults and security rules, plus individual named skill workflows	Read	The server asks that get_gaia_instructions be called before anything else in every new conversation, and says most failures and merchant-data leakage come from skipping it
Routines	Gaia routines with their runs, plus creation, updates, deletion, manual triggering and enabling or disabling	Read / Write	Routines run automatically on the account; approve creation, triggering and enabling
Views	Ticket-list views and view sections, with creation, partial updates and deletion	Read / Write	Views are shared by the whole team; review updates and deletions
Snippets & Products	Reusable AI Agent snippets and their sources, Shopify products with their AI Agent custom info and metafield definitions, and pending product opportunities	Read / Write	Snippets and product info feed what the AI Agent tells customers; approve status changes and custom info edits
Integrations & Widgets	App Store app details, custom HTTP integrations with their request logs, and sidebar widgets with creation, updates and deletion	Read / Write	HTTP integration logs can carry customer data; widget updates replace the whole template
Audiences	Saved audience segments, the segmentation catalog and aggregate audience sizes	Read	Read-only and aggregate; allow for marketing analysis