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

We track the Arcade 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 Arcade 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 Arcade?

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 Arcade MCP server myself?

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

Connect Arcade MCP to your AI agents, securely.

Q: Is Arcade MCP secure?

The server itself exposes 3 tools, and some of them can write to your Arcade 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 Arcade.

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

Get started for free

Get a demo

Try Arcade MCP now

What your agents can ask once Arcade MCP is connected.

Select the tools your AI agents can call

3 tools from the Arcade MCP schema.

Showing 3 tools

Arcade_SelectToolsreadTakes one or more task descriptions and returns the tools that match each task most closely, with their full input schemas. Returns up to five tools per task, ranked by match, plus a query_id for the request, and only tools the gateway allows. A task with no match carries an empty_reason of no_match or unavailable.
Arcade_UseToolhigh riskRuns a tool by name with the inputs you pass, nested inside inputs, and an optional query_id that links the execution to the recommendation that surfaced it. It runs any tool the gateway allows, whether or not Arcade_SelectTools returned it.
Arcade_ListAppsreadLists the apps the user can act on, and whether each is connected, meaning the user can call that app's tools without authorizing first.

What Arcade MCP exposes

Arcade MCP groups its 3 tools into 3 categories, each with a suggested starting policy.

All 3 tools as JSON, ready to paste.

Category	Covers	Access	Suggested policy
Tool selection	Describing one or more tasks, with optional context to help ranking, and getting back the allowed tools that match each one, up to five per task, with full input schemas and a query_id	Read	Read-only and limited to tools the gateway already allows, so it cannot widen access. Matching uses only each tool's name and the first line of its description, which makes the task text the useful thing to log
Tool execution	Running any tool the gateway allows by name, with that tool's own arguments nested inside inputs and an optional query_id linking it to a recommendation	Read / Write	A single entry point for every action on every federated server, so allowing or blocking the tool name decides nothing. Policy has to read tool_name and inputs, and both belong in the log for every call
Apps	Listing the apps the user can act on through the project and whether each is already connected and callable without authorizing first	Read	Read-only, but it reveals which third-party accounts each end user has connected. Low risk on its own and safe to allow for agents that already use the gateway

Connect Arcade 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 Arcade
Find Arcade 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 Arcade MCP or block it. You allow the exact tools each agent can call, one by one.

Arcade MCP	3 TOOLS AVAILABLE
Code review agent	4 tools, read only
Issue triage bot	9 tools
Platform team assistant	all 3 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 Arcade 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.	Arcade MCP3 documented tools listed by an Arcade MCP Gateway with Tool Recommendation on, served at api.arcade.dev/mcp/{gateway-slug} over streamable HTTP. With Tool Recommendation off, or with ?tool_recommendation=false on the URL, the gateway lists every allowed tool from the MCP servers it federates instead, so that toolset is defined per gateway. Arcade_ListApps is documented as the MCP counterpart of GET /v1/apps.
Coding assistants	Connection & credentialsper user	Tool selection1 tool
Internal copilots	Tool access policiesdeny by default	Tool execution1 tool
Workflow automations	Request & response inspectioninline	Apps1 tool
Compatible MCP clients	Activity monitoringalerts and logs