How is this different from connecting GoodData 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 GoodData MCP API changes?

We track the GoodData 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 GoodData 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 GoodData?

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 GoodData MCP server myself?

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

Connect GoodData MCP to your AI agents, securely.

Q: Is GoodData MCP secure?

The server itself exposes 35 tools, and some of them can write to your GoodData 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 GoodData.

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

Get started for free

Get a demo

Try GoodData MCP now

What your agents can ask once GoodData MCP is connected.

Select the tools your AI agents can call

35 tools from the GoodData MCP schema.

Showing 35 tools

search_toolsreadSearches for MCP tools using a natural-language query, ranked over each tool's name, description and parameter descriptions.
call_toolhigh riskCalls a tool returned by search_tools by providing its name and arguments. It is a dispatcher, so it can reach any tool the server exposes, including writes.
get_workspace_inforeadReturns the workspace name, description and organization information.
list_workspace_metricsreadLists metrics with optional RSQL filtering, pagination and include parameters. Passing an entity_id returns a single metric with its full MAQL definition.
list_workspace_attributesreadLists attributes with optional RSQL filtering and pagination, optionally including their labels and datasets.
list_workspace_visualizationsreadLists visualizations with optional RSQL filtering, pagination and single-entity retrieval.
list_workspace_dashboardsreadLists dashboards with optional RSQL filtering and pagination, optionally including their visualizations.
list_usersreadBrowses the organization's users.
list_workspace_usersreadBrowses the users with access to the current workspace.
list_agentsreadLists the organization's AI agents. Requires the AI entitlement.
list_datasourcesreadLists datasources with optional RSQL filtering and pagination. Read-only, with optional single-datasource retrieval.
test_datasourcereadTests datasource connectivity before scanning.
scan_datasourcereadDiscovers the Physical Data Model from the database schemas, reading the warehouse structure directly.
generate_ldmreadGenerates a Logical Data Model from a Physical Data Model produced by a scan.
register_upload_notificationwriteInvalidates related cache after new data is uploaded for a datasource. Organization-scoped and requires the MANAGE permission.
get_workspace_analyticsreadRetrieves the complete analytics model from a workspace, including metrics, dashboards, visualizations, filter contexts, attribute hierarchies, export definitions and dashboard plugins.
create_metricwriteCreates a new governed metric in the workspace. It becomes part of the semantic layer and can be reused by other GoodData capabilities.
create_visualizationwriteCreates a visualization and renders it inline in an MCP Apps host. It requires the experimental MCP Apps feature.
deploy_workspace_analyticshigh riskDeploys a full analytics model to a workspace, replacing the existing analytics model. GoodData flags this with its own warning.
create_memory_itemhigh riskAdds a persistent instruction or piece of context to the AI system prompt for the workspace, up to 255 characters, which later sessions will follow.
ai_searchreadNatural-language search across workspace data with AI-generated insights.
ai_chatwriteA conversational assistant that routes between semantic search over existing dashboards, metrics and visualizations, generating new visualizations from a description, and general analytics questions, with thread continuity across turns.
run_key_driver_analysisreadAnalyses what drove a change in a metric.
compute_label_elementsreadRetrieves distinct label or attribute values with pattern filtering, exact filtering, exclusion mode, dependent filters and pagination up to 10,000 per call.
execute_visualizationreadExecutes a saved visualization and returns its computed result so the client can use the data in downstream analysis.
run_exportwriteRuns an on-demand export as dashboard PDF, tabular XLSX, raw execution data, a PPTX slide deck or a widget image, returning a download URL and file metadata. Large files come back as URLs rather than embedded content.
list_notification_channelsreadDiscovers the available delivery channels for email, Slack and webhooks.
list_automationsreadLists all automations in the workspace.
create_metric_alerthigh riskCreates a metric alert of comparison, range or relative type, with a cron schedule, timezone, metric formatting and recipients. External email addresses are permitted when the channel's allowedRecipients setting is EXTERNAL.
update_metric_alerthigh riskUpdates an alert's thresholds, operators, recipients or schedule, or converts it to a different alert type.
update_alert_statehigh riskPauses or resumes alerts by setting a target state. It accepts either a single automation ID or a list, so one call can silence many alerts at once.
create_scheduled_exporthigh riskCreates a recurring export of a dashboard or visualization, which then delivers on a schedule without further prompting.
get_maql_guidereadReturns MAQL syntax rules, function explanations and examples, so a client can write metric syntax that is not part of general model training data.
list_knowledge_topicsreadLists the available knowledge topics: gooddata, maql, dashboards, visualizations, datasets and ldm-schema.
get_knowledge_topicreadRetrieves the full content of a specific knowledge topic.

What GoodData MCP exposes

GoodData MCP groups its 35 tools into 10 categories, each with a suggested starting policy.

All 35 tools as JSON, ready to paste.

Category	Covers	Access	Suggested policy
Automations & Alerts	Listing the delivery channels available for email, Slack and webhooks, listing the workspace's automations, creating metric alerts of comparison, range or relative type with a cron schedule and recipients, updating an alert's thresholds, operators, recipients or schedule, pausing or resuming one or many alerts, and creating a recurring export	Read / Write	Alerts send to real recipients, including external email addresses when the channel's allowedRecipients permits it, so creating one is an outbound message on a schedule. update_alert_state pauses monitoring without deleting anything and accepts a list of IDs, so one call can silence many alerts
Workspace Metadata	The workspace name, description and organization, and listing metrics, attributes, visualizations and dashboards with RSQL filtering, pagination, single-entity retrieval and an include parameter that pulls related facts, labels, datasets or visualizations in one call	Read	Read-only and bounded by the user's permissions. Retrieving a single metric returns its full MAQL definition, so the output describes how the organization calculates its numbers rather than just naming them
Datasources & Data Modeling	Listing datasources with filtering and single-datasource retrieval, testing connectivity before a scan, discovering the physical data model from database schemas, generating a logical data model from that physical model, and invalidating cache after a data upload	Read / Write	These are organization-scoped and may need elevated permissions such as MANAGE. scan_datasource reads the warehouse schema itself, so the result is a structural map of the underlying database
Analytics Authoring	Retrieving the complete analytics model with its metrics, dashboards, visualizations, filter contexts, attribute hierarchies, export definitions and dashboard plugins, creating a governed metric, creating a visualization rendered inline, and deploying a full analytics model	Read / Write	deploy_workspace_analytics replaces the existing analytics model in the workspace, which GoodData flags as a warning in its own documentation. create_metric adds to the semantic layer, so other capabilities will reuse whatever it defines
Execution & Export	Retrieving distinct attribute values with pattern, exact and exclusion filtering up to 10,000 at a time, executing a saved visualization and returning its computed result, analysing what drove a change in a metric, and running an on-demand export as PDF, XLSX, raw data, PPTX slide deck or a widget image	Read / Write	run_export returns a download URL for large files rather than embedding them, so the data moves outside the server and is invisible to anything watching tool calls alone. It uses existing export permissions and errors when they are missing
Knowledge Base	MAQL syntax rules, function explanations and examples, the list of available knowledge topics covering GoodData, MAQL, dashboards, visualizations, datasets and the LDM schema, and the full content of one topic	Read	Read-only vendor documentation, safe to allow freely. The same topics are also exposed as MCP resources for clients that support resource querying
Users & Agents	Browsing the organization's users, the users with access to the current workspace, and the organization's AI agents	Read	Read-only. The two user lists together describe who can see what, which is an access map worth treating as internal. Listing agents requires the AI entitlement
Tool Discovery	Searching for MCP tools with a natural-language query ranked over tool names, descriptions and parameter descriptions, and calling a tool returned by that search by name and arguments	Read / Write	call_tool is a dispatcher that can invoke anything the server exposes, so its risk is whatever it is pointed at rather than its own. Gate on the inner tool name in the arguments, not on call_tool itself
AI Search & Chat	Natural-language search across workspace data with AI-generated insights, and a conversational assistant that routes between semantic search over existing dashboards and metrics, generating new visualizations from a description, and general analytics questions, with thread continuity across turns	Read / Write	ai_chat can create visualizations as part of answering, so a question can produce an object. Its routing decides that itself, which makes the read and write distinction depend on the prompt rather than the tool
Memory	Adding a persistent instruction or piece of context to the AI system prompt for the workspace, up to 255 characters	Write	This writes standing instructions into the system prompt that other people's sessions will then follow. Treat it as a configuration change to the assistant rather than as a note, and require a named approval for every call

Connect GoodData MCP to your AI agent

Four steps, under ten minutes.

01
Sign in to NeuralTrust
Open the NeuralTrust platform, sign in and go to Getting started, then Connect your tools.
02
Choose your agent
When asked which agent you use, pick Claude Code, Codex, Cursor, Windsurf, Copilot or your own code.
03
Connect GoodData
Find GoodData in the tool list and connect it.
04
Connect your agent
Follow the steps on screen to add the config to your agent, then send the test prompt.

Permissions down to the single tool

You do not allow GoodData MCP or block it. You allow the exact tools each agent can call, one by one.

GoodData MCP	35 TOOLS AVAILABLE
Code review agent	4 tools, read only
Issue triage bot	9 tools
Platform team assistant	all 35 tools

Tool by toolConnecting the server is not the unit of access. Each tool is granted on its own.
Default denyA tool you have not enabled is never exposed to the agent.
Approval on writesTools that change or delete things can require a human to confirm.

Is GoodData MCP secure?

One endpoint instead of raw credentials. Every call is checked against policy and logged.

AI agentsConnect through your MCP client configuration.	NeuralTrustApply controls based on your configuration.	GoodData MCP35 named tools at a workspace-scoped endpoint under /api/v1/actions/workspaces/<id>/ai/mcp, on MCP protocol 2025-11-25. Tool search is enabled by default and cannot be disabled, so tools/list returns only search_tools and call_tool while every other tool stays callable by exact name. The server also generates MCP tools from GoodData v1 metadata API operations: every workspace-scoped GET plus ten allowlisted create entities, with no PUT, PATCH, DELETE, action, or organization-scoped endpoints. Four authentication modes, of which only a shared API token does not run as the end user.
Coding assistants	Connection & credentialsper user	Automations & Alerts6 tools
Internal copilots	Tool access policiesdeny by default	Workspace Metadata5 tools
Workflow automations	Request & response inspectioninline	Datasources & Data Modeling5 tools
Compatible MCP clients	Activity monitoringalerts and logs	All other tools19 tools