Is there an MCP server for Adobe Target?

Yes. Agent Planners runs a hosted MCP server at https://www.agentplanners.com/api/mcp that gives Claude, ChatGPT, Codex and other MCP clients 6 Adobe Target read tools and 3 write tools. Sign in with OAuth or an API key. Writes follow your approval settings; changes you can't take back wait for a person unless an admin exempts a key.

MCP

Adobe Target MCP server: activities, lift and offers in Claude, Codex or Cursor

TL;DR

Adobe Target reaches Claude, Codex or Cursor through the Agent Planners MCP endpoint with 6 reads and 3 writes of its own: activities and their lift by experience, offers, audiences and locations, and approval-gated offers, audiences and saved activities.

What already exists for Adobe Target

Adobe hosts an Adobe Target MCP server (targetmcp.adobe.io), in public beta and read-only during the beta as of November 2026, signed in with your Adobe ID.

Under any layer sits the Target Admin API on mc.adobe.io with your tenant in the path — and a different media type per resource.

What this setup adds

Adobe's hosted serverAgent Planners MCP
WritesRead-only during the betaJSON offers, audiences, saved activities and state changes under the key's write mode
Making an activity live—Held for a person on every key mode
Target toolsThe server's own6 reads and 3 writes, plus two shared connection reads
Media typesHandled by the serverSet per path (offers v2, activities and audiences v3, the rest v1)

How to configure it

  1. 1In Adobe Developer Console add the Adobe Target API to a project with an OAuth Server-to-Server credential, and put the credential on the product profile it should use.
  2. 2On Integrations open the Adobe Target card, press Paste JSON (or Manual entry) and Test connection. The card requests only this product's scopes: openid, AdobeID, target_sdk, additional_info.projectedProductContext, read_organizations, additional_info.roles.
  3. 3Create a key in API & MCP: tools:read for reading, add tools:write for changes. To keep a key to Adobe, choose Only these platforms → Adobe.
  4. 4Add the server to your client with one of the snippets below, run describe_permissions, then adobe__get_connection and adobe__target_list_activities.
bash
# Claude Code
claude mcp add --transport http agent-planners https://www.agentplanners.com/api/mcp \
  --header "Authorization: Bearer $AGENTPLANNERS_API_KEY"

# Codex (signs in with OAuth in the browser)
codex mcp add agent-planners --url https://www.agentplanners.com/api/mcp

# Cursor (~/.cursor/mcp.json)
{ "mcpServers": { "agent-planners": { "url": "https://www.agentplanners.com/api/mcp", "headers": { "Authorization": "Bearer ${env:AGENTPLANNERS_API_KEY}" } } } }
The card also needs your tenant — the name after @ in your Target address; it suggests and confirms it. Every client's setup: MCP setup guide.

The Adobe Target tools

ToolKindWhat it does
target_list_activitiesReadTarget activities (GET activities): id, name, type (ab | xt | …), state (approved = live | deactivated | saved | archived), priority, startsAt / endsAt, modifiedAt.
target_get_activityReadONE activity with its locations, options, experiences and metrics (GET activities/ab/{id} or activities/xt/{id}).
target_list_offersReadOffers (GET offers): id, name, type (content | json …), modifiedAt.
target_list_audiencesReadTarget audiences (GET audiences): id, name, description, origin, modifiedAt.
target_activity_reportReadAn activity's performance report (GET activities/ab/{id}/report/performance or xt): visitors, visits, conversions and lift per experience.
target_getReadANY Target Admin API GET under the tenant (path relative to /{tenant}/target/: activities[/ab|xt/{id}], offers[/content|json/{id}], audiences[/{id}], properties, mboxes, environments, reports …).
target_set_activity_stateWrite — always a personChange an activity's state (PUT activities/state with the id): { accountId?, activityId, state: approved | deactivated | saved }. ★ approved (the activity goes LIVE on the site) is PERSON-GATED on every lane.
target_create_offerWriteCreate an offer (POST offers/content or offers/json): { accountId?, name, type: html | json, content (an HTML fragment, or a JSON object / string) }. ★ An HTML offer is code that runs on the site — PERSON-GATED.
target_mutateWriteAny other Target Admin API write under the tenant (method POST | PUT | DELETE, path relative to /{tenant}/target/, body = Adobe's own fields; contentType? defaults to application/vnd.adobe.target.v3+json): activities …
Each Adobe product is its own connection and its own MCP offering. A workspace that has connected Adobe Target lists and finds its tools; one that has not does not list them at all, and describe_tool on one of them says to connect the Adobe Target card. A key narrowed to the Adobe platform reaches only the Adobe products the workspace has connected; adobe__get_connection shows each one.

What waits for a person, whatever the key allows

Ordinary writes — a JSON offer, an audience, an activity created saved, deactivating an activity — run inside the workspace's guardrails or queue for an admin, as the key was minted. These are held for a person on every key mode:

  • An activity made live (state approved), directly or inside a gateway body.
  • Any HTML or content offer — code that runs on the site.
  • Any delete.
Refused outright: the Target batch endpoint (each change is its own call).

What we measured

  • Offers accept v2 only, activities and audiences v3, mboxes / environments / properties v1 — a wrong one answered 415.
  • The tenant was found from the organization's name with at most four probe reads; a wrong tenant answers 403 502061.
  • An HTML offer created, changed, read back and deleted; an audience created and deleted; an A/B activity created saved, deactivated and deleted — never made live.
More on Adobe Target: what the agent does · setup and every tool · MCP server · AI agent guide · playbook. The other Adobe products, each its own connection: Adobe Analytics · Adobe Experience Platform · Adobe Launch · All four together: Adobe.

Frequently asked questions

Is there an official Adobe Target MCP server?
Yes: Adobe hosts one (targetmcp.adobe.io), in public beta and read-only during the beta as of November 2026. Agent Planners also changes Target — offers, audiences, activity state — with making an activity live always held for a person.
Can an MCP client make a Target activity live?
It can ask; the change is held for a person whatever the key's write mode.
How do I connect Claude to Adobe Target?
Connect Adobe Target on its card on Integrations (with your tenant), create an Agent Planners key, add the MCP server to Claude, and call adobe__target_list_activities.
Put a human-approved agent on your ad ops

Start free — 2,500 credits a month, no credit card. Reads run at once; by default every write waits for your approval.

Related reading