The server itself exposes 28 tools, and some of them can write to your Front 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 Front.

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

We track the Front 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 Front 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 Front?

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 Front MCP server myself?

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

Connect Front MCP to your AI agents, securely.

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

Get started for free

Get a demo

Try Front MCP now

What your agents can ask once Front MCP is connected.

Select the tools your AI agents can call

28 tools from the Front MCP schema.

Showing 28 tools

search_conversationsreadFull-text and filter search across conversations. Scope defaults to my_conversations; all_inboxes is required for unassigned conversations to appear. Filters by inbox, teammate, team, tags, status and date range.
read_conversationreadReturns the full timeline of a conversation: subject, message content, internal comments, tag and assignment changes, status transitions, rule actions and other metadata.
create_draftwriteCreates a draft reply on an existing conversation, or a new outbound conversation. Does not send.
update_draftwriteUpdates the body, subject or recipients of an existing draft, using a version token for conflict detection. Can claim authorship from another teammate with takeOver.
delete_draftwriteDiscards an unsent draft owned by the authenticated teammate. Owner-only.
assign_conversationwriteAssigns a conversation to a teammate, or unassigns it.
move_conversationwriteMoves a conversation to another inbox. Replaces the current inbox association rather than adding to it.
send_messagehigh riskSends a draft created via create_draft. Requires the send scope and carries destructiveHint. The draft must be in compose state, owned by the authenticated teammate, and have at least one recipient.
update_conversation_statushigh riskArchives or reopens a conversation and can set a ticket status. Carries destructiveHint. If the requester is the assignee the change is global, otherwise only their own view changes.
read_messagereadFetches a single message by ID with full content: body with quoted replies stripped, recipients, attachments, author, draft status and any delivery error.
get_attachmentreadGets a single attachment on a message or comment, returning filename, content type, size and a short-lived download URL.
search_message_templatesreadSearches the message templates (saved replies) the authenticated user can access.
add_commentwritePosts an internal comment on a conversation, not visible to customers. Supports @mentions auto-resolved to teammate names.
create_taskwriteCreates a task with subject, assignee, first comment, description, due date, inbox, source conversation, tags, additional teammates and custom fields. The returned ID is a conversation ID.
update_taskwriteUpdates a task's subject, description, due date, status, ticket status, snooze time, associated accounts and custom fields. Every task state change goes through this tool.
list_tagsreadLists all tags in the workspace including hierarchy, filterable by name, inbox, list visibility and parent tag.
tag_conversationwriteAdds or removes tags on a conversation. Idempotent and safe to call repeatedly.
search_contactsreadSearches contacts by name, email or handle.
read_contactreadReturns a full contact record including handles and linked accounts.
search_accountsreadSearches accounts (companies) by name.
read_accountreadReturns the full account (company) record.
get_my_identityreadReturns the calling agent's own identity: public ID, name, alias and whether the caller is human. Takes no arguments.
list_inboxesreadLists inboxes accessible to the authenticated user, filterable by name, access mode, AI, ticketing and business-hours settings, teammates and groups.
list_channelsreadLists channels accessible to the authenticated user, filterable by name, inbox, type and account status. Use it to find channel IDs before calling tools that need one.
list_teammatesreadLists all teammates and their roles in the workspace, filterable by name and account status.
list_teamsreadLists all teams in the workspace.
list_statusesreadLists the company's ticket statuses. Returns empty when ticketing is not enabled.
list_draftsreadLists in-flight draft messages authored by the authenticated teammate.

What Front MCP exposes

Front MCP groups its 28 tools into 7 categories, each with a suggested starting policy.

All 28 tools as JSON, ready to paste.

Category	Covers	Access	Suggested policy
Conversations	Full-text and filtered conversation search, the complete conversation timeline, draft creation, editing and deletion, sending, archiving and reopening, assignment and moving between inboxes	Read / Write	send_message needs the send scope and is the only tool that reaches a customer; drafting is a separate step, so a review copilot can have read and write without send
Organization	The calling agent's own identity, inboxes, channels, teammates, teams, ticket statuses and the teammate's in-flight drafts	Read	Discovery tools; list_channels is the prerequisite for any tool that needs a channel ID
Contacts & Accounts	Contact search by name, email or handle, full contact records with handles and linked accounts, and company account search and records	Read	Contact records are customer personal data; restrict and log
Messages	Reading a single message with full body, recipients, attachments and delivery errors, and fetching an attachment's metadata with a short-lived download URL	Read	get_attachment returns a download URL to customer-supplied files; treat those files as untrusted
Tasks	Creating tasks and updating their subject, description, due date, status, snooze, accounts and custom fields. Task IDs are conversation IDs	Read / Write	Task state changes go through update_task, never update_conversation_status
Tags	Listing workspace tags including their hierarchy, and adding or removing tags on a conversation	Read / Write	Tagging is idempotent and low risk, but tags often drive routing rules
Comments	Posting an internal comment on a conversation, with @mentions resolved automatically	Read / Write	Internal only and not visible to customers, but @mentions notify real teammates

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

Front MCP	28 TOOLS AVAILABLE
Code review agent	4 tools, read only
Issue triage bot	9 tools
Platform team assistant	all 28 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 Front 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.	Front MCP28 documented tools across three OAuth scopes: read, write and send.
Coding assistants	Connection & credentialsper user	Conversations9 tools
Internal copilots	Tool access policiesdeny by default	Organization7 tools
Workflow automations	Request & response inspectioninline	Contacts & Accounts4 tools
Compatible MCP clients	Activity monitoringalerts and logs	All other tools8 tools