MCP server

Connect Claude, ChatGPT, Cursor or your own agent to Pluto Ads through the Model Context Protocol.

The server offers every Pluto Ads action as a tool. Each tool runs the same action as the app and the Public API, with the same permissions, idempotency and audit log. For connecting from the app's side, and what your team sees, read Connected agents.

Server URL

Server URL
https://api.plutoads.ai/api/v1/mcp

SettingsAPI (Settings, then API) shows it under Connect an MCP client, with the setup for your client. The server speaks Streamable HTTP and offers tools only.

Connect a client

Clients sign in with OAuth: add the server, and your client opens Pluto Ads in the browser to sign in. There's no key to copy.

ClientSetup
ClaudeIn Claude, open CustomizeConnectors (Customize, then Connectors), select Add custom connector and paste the server URL. See Claude's guide.
ChatGPTTurn on Developer mode in SettingsSecurity and login (Settings, then Security and login), then create an app under Plugins with the server URL and OAuth. See ChatGPT's guide.
CursorAdd the config to ~/.cursor/mcp.json. See Cursor's guide.
Claude CodeRun the command, then choose Authenticate under /mcp. See Claude Code's guide.
VS CodeRun MCP: Open User Configuration from the Command Palette and add the config. See VS Code's guide.
CodexRun the command in your terminal. See Codex's guide.
{
  "mcpServers": {
    "plutoads": {
      "url": "https://api.plutoads.ai/api/v1/mcp"
    }
  }
}

Other clients point a Streamable HTTP client at the server URL. A client that can't sign in sends an API key as Authorization: Bearer <key> instead:

import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";

const transport = new StreamableHTTPClientTransport(new URL("https://api.plutoads.ai/api/v1/mcp"), {
  requestInit: { headers: { Authorization: `Bearer ${process.env.PLUTO_KEY}` } },
});
const client = new Client({ name: "brevanne-reports", version: "1.0.0" });
await client.connect(transport);
const context = await client.callTool({ name: "get_context", arguments: {} });

Once connected, ask the agent to call get_context. It answers with who the agent is, its workspace and what it may do now. capabilities lists each thing it may do, with what it allows. role is the role of the person it acts for. notGranted lists what that person may do that this connection may not:

get_context
{
  "actor": {
    "id": "01a0d9f2-6c1e-7b3d-9a41-5e2f8c7d1a90",
    "kind": "agent",
    "name": "Claude via Sarah Mitchell",
    "capabilities": [
      { "id": "ads:read", "allows": "View ads, launches, analytics and spend" },
      { "id": "reviews:write", "allows": "Comment, review and approve ads" },
      { "id": "launches:write", "allows": "Create and edit launches and drafts" }
    ],
    "permissions": ["read", "review"],
    "userId": "01a0d7e5-3b7c-7e52-8d14-2a6f9b0c4e17"
  },
  "role": { "workspace": "manager", "workspaceRoleName": "Manager", "organization": "member" },
  "notGranted": [{ "id": "ads:publish", "allows": "Publish and launch ads to the platform" }],
  "workspace": { "id": "01a0d7e5-acb8-7176-a698-b6947a8016fe", "name": "Brevanne", "version": 4 }
}

permissions lists the older scope names whose whole set the credential holds, for clients that read them.

Sign-in and scopes

The server is an OAuth 2.1 protected resource. A request without a credential gets 401 with a WWW-Authenticate header that points to its metadata:

Metadata
curl -s https://api.plutoads.ai/api/v1/.well-known/oauth-protected-resource

The metadata names the authorization server. Clients read it, register and sign the person in with PKCE, so nothing needs configuring by hand. The same metadata is at the origin root, https://api.plutoads.ai/.well-known/oauth-protected-resource, for clients that look there.

When a client signs in, the browser opens Connect to Pluto Ads. The person chooses the workspace and the access level. The level becomes the token's scopes: read, review, write, publish, admin and organization, as in Scopes. The token is short-lived, its audience is the server URL, and it never does more than the person's current role allows.

The server also accepts:

  • API keys, as a Bearer token. A publishable key only calls read tools.
  • Agent credentials, exchanged for a token. See Agent credentials.

Every tool names the capability it needs in _meta["ai.plutoads/capability"] and in its description. The scope that holds it is in _meta["ai.plutoads/permission"]. The server checks the capability on every call, against the person's current role and the credential's scopes. tools/list returns only the tools the credential can use now, so the list changes when the person's role or the credential does. A tool called from an older list is refused, and the refusal names the capability it needs. A tool only a person signed in to Pluto Ads completes, such as open_billing_portal, is never in a credential's list, whatever it holds.

get_context says what the credential holds and why: actor.capabilities (what it may do now), role (the person's workspace and organization roles), notGranted (what the person may do that this credential may not: ask them for a credential that includes it) and credential (for a key or agent: the person it acts for, when it expires, the addresses it works from and what it reaches). Where the organization lets members' agents only view (credential policy), a member's agent holds the view capabilities only, and its tool list says so.

Approvals

Tools that spend money or change what runs in Meta or LinkedIn or X need the publish scope. The server tells agents to confirm these with the person first, and the confirmation happens in the agent's own conversation:

  1. The agent calls get_publish_preview, which returns the launch's launchVersion, adAccountId and budgetSummary.
  2. It shows the person that summary and waits for a yes.
  3. It calls publish_launch with those values as its approval. If the launch, the ad account or a budget changed since, the server refuses, and the agent has to ask again.

Pluto Ads can't see the conversation, so give publish only to agents you trust to ask. Live changes like set_budget and set_delivery_status also say to confirm with the person. delete_social_comment refuses to run without confirm: true.

For several live changes at once, an agent can call propose_changes instead. The person reviews that plan in the Pluto Ads Inbox, and nothing changes until they approve it there. Agents can't approve plans, loop runs or the in-app Agent's proposals: those are always a person's, in the app.

Tools carry MCP annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint), so a client can run read-only tools without asking and confirm the rest.

How tools behave

  • Start with get_context. Then refer to everything by ID. Names aren't unique.

  • One ad platform per call. Tools that work on more than Meta take platform: meta (the default) or linkedin, or x for X. list_platforms lists the platforms the workspace can use and what Pluto Ads can do on each, with the settings each platform's loops take and what saving refuses there (loops.rules, loops.refusedBlocks), and get_platform_spec gives a platform's launch settings, options and rules in its own words. validate_loop checks a loop exactly as create_loop, update_loop and turning it on do.

  • X's data stays in Pluto Ads. X's terms keep what Pluto Ads reads from X in its own app until X agrees in writing that it may leave. On X, a tool answers with IDs, your launches and what Pluto Ads did, and each field that holds what X reports (a name, a status, a figure, a lead's answers, a reply's text) is null with withheld: true beside it. See What stays in Pluto Ads.

  • Writes are commands. Every write tool takes an optional idempotencyKey. Reuse it when you retry, and the stored result comes back instead of a second change.

  • Batch tools take items[] and return an outcome per item: accepted, rejected or conflict. Resend only what failed.

  • Updates take expectedVersion from your last read. See Versions and conflicts.

  • Long work returns a job. Follow it with wait_for_job, which streams progress when you pass a progressToken, then list_job_items for what failed. cancel_job needs what the job's work needs: publish for publishing and changes to what runs, and admin for a spending limit.

  • Media never travels through MCP. import_media fetches up to 1,000 HTTPS URLs per job, and create_media_uploads returns URLs to upload files to. For hundreds or thousands of files, start with prepare_bulk_upload: it names a folder and the ways to fill it (an uploader for agents with a shell, a link for a person to drop files, URL import), and create_ads_from_media then makes an ad per file. See Bulk upload.

  • Looking at creatives. view_media and view_ads return images your model can see: a still of an image, or frames of a video.

  • Profit comes from Pluto Profit, read only, one store per workspace. connect_pluto_profit returns a link for a person to open and approve. Money is decimal strings, in the store's currency for store and product figures and in the ad account's for ad estimates, and a value Pluto Profit can't compute is null, never zero. get_ad_profitability is always an estimate.

  • Results have a one-line summary, the JSON, and structuredContent that matches the tool's output schema. Large reads are compact by default: get_launch returns each ad's copy rather than its merged settings and pages its ads 50 at a time (limit, after, adSetId); verbose: true returns the whole document. Batch writes that change ads, ad sets, launches or media (update_ads, move_ads, edit_ad_sets, apply_bulk_edit, revert_changeset, delete_media and the other media batch tools) return each item's outcome with its ID, version, name and status; pass verbose: true for the whole entity. rank_ads returns the ranked metric with the core ones (metrics for others); get_insights and get_launch_results take metrics to keep only those, and give byCurrency only when amounts span several currencies. search_creative_ads pages 10 ads and list_media 20 items at a time unless limit says otherwise.

  • Performance. get_insights reads what the Analytics dashboard shows, with the same filters as GET /analytics/insights: segment filters (age, gender, country, placement, hour and more), breakdown by launch, campaign, ad set, ad or any segment, attribution including 28d_click, and metrics that take any action Meta counts as action:<type>, action_value:<type> or cost_per_action:<type>. rank_ads ranks by any of those metrics, within a segment too, and its rows carry Meta's quality, engagement rate and conversion rate rankings, as do get_creative_performance's. Segment numbers reach back 13 months and use Meta's default attribution setting (each ad set's own, as in Ads Manager); reach and unique clicks come from Meta for exactly the period asked. By region, product and the action breakdowns Meta reports no website or app conversions: the answer's note says so, and they're absent, never 0. A metric marked incomplete reads fields older days are still being read again with (coverage.history.fieldsCompleteFrom). With platform: "linkedin" every LinkedIn metric is there (engagements, qualified leads, open rate, viral figures and the rest, ranked by rank_ads too) and the segments are conversionAction, placement, device, servingLocation and carouselCard; LinkedIn's demographic segments (jobTitle, jobFunction, seniority, industry, companySize, company, country, region, county) are refused, since LinkedIn doesn't allow its demographic data as input to AI. The API serves them.

  • Exactly the numbers you need. get_insights and get_launch_results check metrics against the platform's catalogue: an ID it doesn't report is refused with the closest IDs it has. Each metric asked for (or every one, with definitions: true) comes with its definition: name, unit, how it adds up (aggregation: sum, ratio, weighted or window) and summable. Reach, frequency and unique clicks are window metrics: breakdown rows never add up to their total. With a breakdown, sort (a metric, descending by default), limit (up to 1000) and cursor (the previous breakdown.nextCursor) page the rows, and breakdown.rowsTotal counts them all. Filter by name with campaignNames, adSetNames, adAccountNames and launchNames when you have no IDs, by status with statuses and by Pluto's labels with labelIds or labelNames (every figure then covers only those ads). thenBy splits each breakdown row again by day, week, month or a segment (ad by day, campaign by country): each row's by.rows add up to it, and window metrics are never in them. platforms (or ["all"]) answers several platforms in one call, each on its own under byPlatform and never added together; withheldPlatforms and refused name any that can't give numbers there.

  • Results kept in Pluto. Where a platform's terms keep its results inside Pluto's own app, a tool's answer holds withheldData, one entry per platform: reason (platform_terms), a message to act on, appUrl (the workspace's Analytics on that platform), dataExists: true, retryable: false and stillAvailable. The results exist and are complete in Pluto; the nulls are not missing data or zero, and calling again or another tool gives the same answer. Tell the person to open appUrl and keep working with the IDs, your launches and every other platform's data. The server's instructions say the same when you connect.

  • The audit log. list_audit_log reads the workspace's log, or with organizationId the organization's (for its owners and admins). Each entry's detail has what changed where it says more than its summary, and requestId the request it came from. commandId finds every entry of one command. export_audit_log writes the same entries, with the same filters, to a CSV file in the background: follow its jobId with wait_for_job, then get_audit_log_export gives downloadUrl (15 minutes; the file is kept for 7 days). The CSV's columns are in the Public API.

  • Versions. Changes send the expectedVersion they read: search_ads and get_launch return each ad's version. A bulk edit records a changeset, and revert_changeset puts back what it changed even after other changes since (a move), as long as the fields it wrote still hold its values. list_bulk_edits lists your bulk edits with canUndo, so a lost changesetId can be found again.

  • Readiness. get_launch_review and the app's checklist flag an ad whose file failed processing or was removed (publishing would refuse it); files still uploading or processing don't block, publishing waits for them. Checklist items carry a fix link into the app, like pluto:launch/<launchId>#adAccountId.

  • X. Every tool takes "platform": "x" where it takes a platform. X's terms keep what Pluto reads from X inside Pluto, so in a tool's answer a field holding what X reports (a name, a status, a figure, a lead's answers, a reply's text) is null and withheld beside it is true; withheldData and the summary say why and link the workspace's Analytics on X, where the person sees them. IDs, your launches and drafts, and the outcome of what you asked Pluto to do answer as always. update_lead_form and archive_lead_form change and archive an X lead form, get_insight_segments splits a campaign's results by age, gender, platform, region or metro, estimate_platform_audience asks X for an audience's size and get_ad_preview for X's preview of an ad: the last three answer withheld: true here, and X isn't asked. manage_platform_conversion_event creates, changes and deletes an X ad account's conversion events; the app's X page lists them with the windows each counts in and the code that fires it on your site. list_platform_audiences lists an ad account's custom audiences and do-not-reach list by ID, and manage_platform_audience makes a customer list from people's emails, phone numbers, usernames, X user IDs or device IDs, adds and removes people, renames and deletes it: Pluto hashes every value before it reaches X and keeps none of them.

Errors and limits

A problem with the request is a tool result with isError: true, so the model can read it and act. It holds a message and the same code and fields as the API errors.

Tool error
{
  "content": [
    { "type": "text", "text": "You can view this workspace but not change it. It needs the `launches:write` capability (create and edit launches and drafts). Ask an admin to give this credential the access it needs; it never does more than its person can." },
    { "type": "text", "text": "{\"code\":\"forbidden\",\"message\":\"You can view this workspace but not change it.\",\"missingCapability\":\"launches:write\",\"missingCapabilityText\":\"Create and edit launches and drafts\"}" }
  ],
  "isError": true
}

The message says what to do next:

  • Arguments that don't fit name the argument (`items[2].adId`) and, for an ID or a time, what goes there.
  • A refusal for the credential's access says which access it needs. Tools that are a person's own (their inbox, Agent chats, views, skills and preferences) need a client signed in with OAuth as that person; API keys and machine-to-machine clients can't use them.
  • When Pluto or the platform is limiting how often something runs, the message gives the seconds to wait (retryAfter in the JSON).

Credential and transport problems are JSON-RPC errors:

HTTPCodeMeaning
401-31001No credential, or it's invalid, expired or revoked. Sign in again.
403-31003The credential can't use this workspace, or the request's origin isn't allowed.
429-31029Over the rate limit. data.retryAfter and Retry-After give the seconds.
503-31503Sign-in is unavailable. Retry shortly.
400-32020, -32022Headers that don't match the body, or an unsupported protocol version.
200-32603Our fault. Retry with the same idempotencyKey.

MCP and the REST API share the same rate limits: 1,000 requests per 10 seconds per credential, and 3,000 for a workspace's keys and agents together.

Protocol versions

The server has no sessions: each request stands alone, and GET or DELETE on the URL answers 405. A request is one JSON-RPC message, and batches aren't supported.

VersionsHow it works
2026-07-28Stateless. Each request carries the protocol version and client capabilities in _meta, plus the MCP-Protocol-Version, Mcp-Method and, for tools/call, Mcp-Name headers.
2025-11-25, 2025-06-18, 2025-03-26initialize agrees on a version: yours if we support it, otherwise 2025-11-25.

Tool catalogue

Every tool the server offers, 448 in all, by area. Access is the permission a tool needs. tools/list has each tool's full description and input and output schemas.

Workspace. Who you are, the workspace, its people and who is here.

ToolWhat it doesAccess
get_contextGet your identity and permissionsread
list_workspacesList workspacesread
create_workspaceCreate a workspaceorganization
list_membersList membersread
invite_membersInvite membersadmin
list_invitationsList invitationsread
revoke_invitationRevoke invitationadmin
edit_membersEdit membersadmin
rename_workspaceRename workspaceadmin
update_workspace_settingsUpdate workspace settingsadmin
create_workspace_logo_uploadStart workspace logo uploadadmin
set_workspace_logoSet or remove workspace logoadmin
list_webhook_eventsWebhook eventsread
get_presenceWho is hereread
list_audit_logList the audit logadmin
export_audit_logExport the audit log as CSVadmin
get_audit_log_exportGet an audit log exportadmin

Launches. Launches, ad sets and ads, bulk edits and launch specs.

ToolWhat it doesAccess
list_launchesList launchesread
search_adsSearch ads across launchesread
get_launchGet launchread
get_launch_reviewWhat's missing before publishingread
create_launchesCreate launcheswrite
update_launchesUpdate launcheswrite
archive_launchArchive or delete launchwrite
archive_launchesArchive or restore launcheswrite
duplicate_launchDuplicate launchwrite
get_launch_statsLaunch overviewread
edit_ad_setsEdit ad setswrite
create_adsCreate adswrite
create_ads_from_mediaCreate ads from mediawrite
update_adsUpdate adsreview
duplicate_ad_setDuplicate an ad setwrite
split_ad_setSplit an ad setwrite
move_adsMove adswrite
duplicate_adsDuplicate adswrite
delete_adsDelete adswrite
list_archivedArchived ads and ad setsread
restore_archivedRestore archived ads or ad setswrite
preview_bulk_editPreview bulk editwrite
apply_bulk_editApply bulk editwrite
list_bulk_editsList bulk editsread
get_changesetGet a bulk editread
revert_changesetRevert a bulk editwrite
validate_launch_specValidate launch specwrite
apply_launch_specApply launch specwrite
list_launch_groupsList launch groupsread
get_launch_groupGet launch groupread
create_launch_groupCreate launch groupwrite
update_launch_groupUpdate launch groupwrite
add_launch_campaignAdd a campaign to a launchwrite
archive_launch_groupArchive or delete launch groupwrite
duplicate_launch_groupDuplicate launch groupwrite
get_launch_group_resultsCompare a creative across campaignsread
get_linked_adsGet linked adsread
update_linked_adsUpdate linked adswrite
unlink_adUnlink adwrite
get_ad_platform_settingsGet an ad's platform settingsread
update_ad_platform_settingsChange an ad's platform settingswrite

Publishing and live changes. Publishing, and changes to what runs in Meta and LinkedIn and X.

ToolWhat it doesAccess
get_publish_previewPrepare to publishread
publish_launchPublish a launchpublish
cancel_publishCancel publishingpublish
get_publicationWhat is liveread
get_revisionsPending changes to live adsread
get_ad_meta_previewMeta's preview of an adread
publish_revisionsPublish pending changespublish
discard_revisionsDiscard pending changeswrite
set_delivery_statusPause or activate live deliverypublish
set_budgetChange a live budgetpublish
set_bidChange a live bidpublish
set_scheduleChange a live ad set's schedulepublish
duplicate_at_metaDuplicate a live campaign, ad set or adpublish
move_at_metaMove a live ad or ad set in Metapublish
delete_at_metaDelete a live ad or ad setpublish
list_keywordsList keywordsread
list_negative_keywordsList negative keywordsread
list_negative_keyword_listsList negative keyword listsread
list_search_termsList search termsread
list_recommendationsList recommendationsread
get_keyword_ideasGet keyword ideasread
get_keyword_metricsGet keyword metricsread
get_keyword_competitionGet keyword competitionread
get_search_term_brandsGet brand search termsread
set_brand_termsSet brand termswrite
forecast_keywordsForecast keywordsread
set_keyword_statusPause or turn on a keywordpublish
add_keywordsAdd keywordspublish
add_negative_keywordsAdd negative keywordspublish
change_keywordsChange keywords, negatives and listspublish
apply_recommendationApply recommendationspublish
dismiss_recommendationDismiss recommendationspublish
remove_campaignRemove a live campaignpublish
get_creative_previewGet the platform's ad previewread
archive_at_platformArchive a live campaign, ad set or adpublish
set_campaign_spend_capChange a live campaign's spend cappublish
search_targetingSearch interests, behaviors and demographicsread
list_audiencesList audiencesread
get_minimum_budgetGet minimum budgetread
set_targetingChange live targetingpublish
estimate_platform_audienceEstimate an audienceread
get_ad_previewPreview ads on their platformread
list_platform_audiencesList a platform's audiencesread
manage_platform_audienceChange a platform's audiencespublish
get_ad_set_estimateEstimate a draft ad setread
list_attachment_choicesList attachment choicesread
list_saved_audiencesList saved audiencesread
manage_saved_audienceChange saved audiencespublish
list_third_party_audiencesList third-party audiencesread

Change sets. Plans of live changes a person approves in the app.

ToolWhat it doesAccess
list_change_setsList plans of live changesread
get_change_setGet a plan of live changesread
propose_changesPropose changes for approvalwrite
revise_change_setSend a plan back with feedbackwrite
edit_change_setEdit a planwrite
undo_changeUndo an applied changepublish

Ad platforms. The Meta and LinkedIn and X connections, their launch settings, synced campaigns, ad sets and ads, and imports.

ToolWhat it doesAccess
get_platform_mirrorGet Keep in syncread
update_platform_mirrorChange Keep in syncadmin
get_meta_integrationGet Meta connectionread
sync_metaSync Metawrite
list_meta_attentionList what Meta needsread
check_meta_attentionCheck a Meta step is donepublish
list_meta_campaignsList Meta campaignsread
list_meta_ad_setsList Meta ad setsread
list_meta_adsList Meta adsread
get_ad_postsGet ad post linksread
get_meta_deliveryGet learning and delivery issuesread
import_from_metaImport from Metawrite
preview_meta_importPreview an import from Metaread
get_import_reportGet an import reportread
switch_meta_ad_accountTurn an ad account on or offadmin
switch_meta_pageTurn a Page on or offadmin
switch_meta_pagesTurn Pages on or offadmin
switch_meta_ad_accountsTurn ad accounts on or offadmin
set_meta_account_pagesSet an ad account's Pagesadmin
set_spending_limitSet an ad account's spending limitadmin
connect_metaConnect Metaadmin
reconnect_metaReconnect Metaadmin
disconnect_metaDisconnect Metaadmin
get_platform_integrationGet a platform connectionread
connect_platformConnect a platformadmin
sync_platformSync a platformwrite
disconnect_platformDisconnect a platformadmin
switch_platform_ad_accountTurn an ad account on or offadmin
switch_platform_ad_accountsTurn ad accounts on or offadmin
switch_platform_pageTurn a Page on or offadmin
switch_platform_pagesTurn Pages on or offadmin
list_platformsList ad platformsread
get_platform_specGet a platform's launch settingsread
list_campaignsList campaignsread
list_ad_setsList ad setsread
list_adsList adsread
search_platform_targetingSearch a platform's targetingread
list_platform_conversionsList an ad account's conversionsread
import_platform_campaignsImport campaigns from an ad platformwrite
get_platform_import_reportGet an import's reportread
list_account_videosList an ad account's videosread
add_account_videoAdd a video to an ad accountwrite
list_conversion_actionsList conversion actionsread
get_conversion_actionGet a conversion action and its tagread
create_conversion_actionCreate a conversion actionpublish
list_account_assetsList an ad account's own assetsread
add_account_assetAdd an asset to an ad accountpublish
remove_account_assetRemove an asset from an ad accountpublish
list_bid_strategiesList portfolio bid strategiesread
create_bid_strategyCreate a portfolio bid strategypublish
update_bid_strategyChange a portfolio bid strategypublish
manage_platform_conversion_eventManage a conversion eventpublish
manage_platform_conversion_goalManage a conversion goalpublish
manage_platform_site_tagManage a site tagpublish
upload_offline_conversionsUpload offline conversionspublish
list_platform_catalogsList product catalogsread
manage_platform_catalogChange product catalogspublish
list_product_setsList product setsread
manage_product_setChange product setspublish
list_catalog_feedsList product feedsread
manage_catalog_feedChange product feedspublish
list_catalog_productsList catalog productsread
manage_catalog_productsChange catalog productspublish
list_catalog_importsList catalog importsread

Results. Ad performance from Meta and LinkedIn and X.

ToolWhat it doesAccess
get_insightsAd performanceread
get_launch_resultsLaunch resultsread
rank_adsRank adsread
refresh_insightsRefresh ad performancewrite
get_platform_reportGet a platform reportread
get_insight_segmentsResults by segmentread
get_delivery_insightsGet a delivery breakdownread
get_reach_overlapGet a reach overlapread

Profit. Profit from Pluto Profit: the store connection, the store's and its products' profit, and an estimate per ad.

ToolWhat it doesAccess
get_pluto_profit_integrationGet Pluto Profit connectionread
disconnect_pluto_profitDisconnect Pluto Profitadmin
connect_pluto_profitConnect Pluto Profitadmin
reconnect_pluto_profitReconnect Pluto Profitadmin
get_profit_summaryStore profitread
get_product_profitabilityProduct profitabilityread
get_ad_profitabilityAd profitability (estimate)read

Research. Competitors' ads in the public Ad Libraries, Reddit threads and comments, web pages, and competitors researched in the workspace: their ads, what Creative Memory found in them and their funnels.

ToolWhat it doesAccess
search_ad_librarySearch Meta's Ad Libraryread
research_ad_libraryRead a public Ad Libraryread
read_web_pageRead a web page in a browserread
search_redditSearch Redditread
read_reddit_threadRead a Reddit threadread
research_competitorResearch a competitor's adswrite
get_competitor_adsRead competitors' adsread
analyze_competitor_funnelAnalyze a competitor's funnelread

Media. The media library, and looking at media and ads.

ToolWhat it doesAccess
list_mediaList mediaread
get_mediaGet mediaread
import_mediaImport media from URLswrite
create_media_uploadsStart media uploadswrite
complete_media_uploadsFinish media uploadswrite
complete_media_uploadFinish a media uploadwrite
prepare_bulk_uploadStart a bulk uploadwrite
update_mediaRename or move mediawrite
move_mediaMove mediawrite
archive_mediaArchive mediawrite
unarchive_mediaUnarchive mediawrite
delete_mediaDelete mediawrite
list_media_foldersList media foldersread
edit_media_foldersEdit media folderswrite
view_mediaLook at mediaread
view_adsLook at adsread

Creative Memory. Search inside your creatives and read what Creative Memory knows about them.

ToolWhat it doesAccess
get_creative_performanceCreative performance evidenceread
compare_variantsCompare creative variantsread
evaluate_creative_conditionCheck a creative conditionread
get_creative_contextCreative compositionread
get_creative_memory_usageCreative Memory statusread
search_creative_memorySearch Creative Memoryread
get_creative_memoryCreative Memory evidenceread
quote_creative_memoryPreview a Creative Memory analysiswrite
request_creative_memoryRead files with Creative Memorywrite
list_creative_memory_requestsCreative Memory requestsread
cancel_creative_memory_requestCancel Creative Memory workwrite
search_creative_adsSearch ads by copy and creativeread
create_creative_library_importIndex historical librarywrite
get_creative_library_importLibrary import progressread
list_creative_library_importsLibrary importsread
list_creative_library_import_itemsLibrary import itemsread
pause_creative_library_importPause library importwrite
resume_creative_library_importResume library importwrite
cancel_creative_library_importCancel library importwrite
retry_creative_library_import_itemRetry library import itemwrite
quote_creative_precisionPreview targeted creative inspectionwrite
request_creative_precisionStart targeted creative inspectionwrite
get_creative_precisionTargeted creative evidenceread
cancel_creative_precisionCancel targeted inspectionwrite
list_creative_monitorsList creative monitorsread
get_creative_monitorRead creative monitorread
list_creative_monitor_itemsList creative monitor itemsread
get_creative_correction_contextReview creative correction sourcesread
correct_creative_evidenceCorrect creative evidencewrite
release_creative_correctionsRelease creative correctionswrite
get_creative_correctionCreative correction historyread
list_creative_correctionsCreative correction revisionsread
record_creative_learning_labelRecord creative learning annotationwrite
get_creative_learning_labelCreative learning annotationread
list_creative_learning_labelsCreative learning annotationsread
register_creative_experimentRegister creative experimentwrite
amend_creative_experimentAmend creative experimentwrite
set_creative_experiment_statusUpdate experiment statuswrite
observe_creative_experimentRecord experiment outcomeswrite
get_creative_experimentRegistered experimentread
list_creative_experimentsCreative experimentsread
list_creative_experiment_outcomesRecorded experiment outcomesread
record_creative_lineageRecord creative relationshipwrite
list_creative_lineageCreative relationshipsread
get_creative_lineageCreative relationship historyread

Review and comments. Comments, review, the Inbox and workflow settings.

ToolWhat it doesAccess
list_commentsList commentsread
add_commentsAdd a review commentreview
create_comment_attachment_uploadStart comment attachment uploadreview
update_commentsEdit, resolve, react to commentsreview
get_ad_activityAd activityread
follow_adSubscribe to an adread
follow_threadSubscribe to a threadread
get_inboxInboxread
update_inboxTriage inboxread
notify_membersNotify peoplewrite
get_workflow_settingsGet workflow settingsread
edit_statusesEdit review statuseswrite
edit_labelsEdit tagswrite
edit_copy_templatesEdit copy templateswrite
edit_country_groupsEdit country groupswrite

Comments on ads and leads. What people post on your ads, and the leads they send.

ToolWhat it doesAccess
list_social_commentsList comments on adsread
mark_social_post_readMark a post's comments readread
set_social_post_stateFollow up or move a post to doneread
get_social_commentGet a comment on an adread
reply_to_social_commentReply to a commentpublish
hide_social_commentHide a commentpublish
hide_social_commentsHide or unhide commentspublish
unhide_social_commentUnhide a commentpublish
delete_social_commentDelete a commentpublish
sync_social_commentsRead comments nowwrite
list_lead_formsList lead formsread
list_leadsList leadsread
get_leadGet a leadwrite
create_lead_formCreate a lead formpublish
sync_leadsRead leads nowwrite
update_lead_formChange a lead formpublish
archive_lead_formArchive a lead formpublish

Loops. Automations: build, test, run and undo, and their folders.

ToolWhat it doesAccess
list_loopsList loopsread
get_loopGet a loopread
list_loop_templatesList loop templatesread
list_loop_skillsList loop skillsread
create_loopCreate a loopwrite
update_loopUpdate a loopwrite
duplicate_loopDuplicate a loopwrite
archive_loopArchive a loopwrite
delete_loopDelete a loopwrite
unarchive_loopUnarchive a loopwrite
validate_loopCheck a loopread
set_loop_enabledTurn a loop on or offwrite
run_loopRun a loop nowwrite
test_run_loopTest a loopwrite
list_loop_runsList a loop's runsread
cancel_loop_runStop a loop runwrite
undo_loop_changesUndo a loop's changeswrite
get_loop_runGet a loop runread
list_loop_foldersList loop foldersread
create_loop_folderCreate a loop folderwrite
rename_loop_folderRename a loop folderwrite
delete_loop_folderDelete a loop folderwrite
move_loopsMove loops to a folderwrite
set_loops_enabledTurn loops on or offwrite

Views. What a person sees on a list page (filters, grouping, ordering and saved views), and where anything is in the app.

ToolWhat it doesAccess
open_viewOpen a viewread
get_viewGet what a list page showsread
set_viewSet what a list page showsread
find_in_appFind in the appread
list_viewsList saved viewsread
save_viewSave a viewread
update_viewUpdate a saved viewread
delete_viewDelete a saved viewread

Agent. The in-app Agent's chats, guidance, skills, models and approvals.

ToolWhat it doesAccess
list_agent_chatsList Agent chatsread
get_agent_chatGet an Agent chatread
rename_agent_chatRename an Agent chatread
delete_agent_chatDelete an Agent chatread
share_agent_chatShare an Agent chatwrite
unshare_agent_chatStop sharing an Agent chatwrite
stop_agent_chatStop an Agent answerread
mark_agent_chat_readMark an Agent chat readread
edit_agent_chat_messageEdit a message in an Agent chatwrite
list_agent_chat_message_versionsList a message's versionsread
select_agent_chat_message_versionSwitch a message's versionwrite
get_agent_guidanceGet Agent guidanceread
set_agent_guidanceSet Agent guidancewrite
list_agent_skillsList Agent skillsread
get_agent_skillGet Agent skillread
create_agent_skillCreate Agent skillwrite
update_agent_skillUpdate Agent skillwrite
delete_agent_skillDelete Agent skillwrite
run_agent_skillRun Agent skillread
list_agent_memoriesList Agent memoryread
get_agent_memoryGet Agent memory entryread
create_agent_memoryRemember for Agentwrite
update_agent_memoryUpdate Agent memory entrywrite
delete_agent_memoryForget Agent memory entrywrite
list_modelsList modelsread
update_model_settingsUpdate model settingsadmin
update_model_visibilityUpdate model visibilityread
get_agent_approval_settingsGet Agent approval settingsread
set_agent_approval_settingsSet Agent approval settingsadmin

Jobs. Long-running work and its items.

ToolWhat it doesAccess
get_jobGet jobread
list_jobsList jobsread
list_job_itemsList job itemsread
cancel_jobCancel jobwrite
retry_jobRetry failed job itemswrite
wait_for_jobWait for jobread

Defaults and templates. What new campaigns and ad sets start with, and campaign and ad set templates.

ToolWhat it doesAccess
get_defaultsGet campaign and ad set defaultsread
set_defaultsSet campaign or ad set defaultswrite
list_setting_templatesList settings templatesread
get_setting_templateGet a settings templateread
create_setting_templateCreate a settings templatewrite
update_setting_templateChange a settings templatewrite
duplicate_setting_templateDuplicate a settings templatewrite
delete_setting_templateDelete a settings templatewrite
apply_setting_templateApply a settings templatewrite

Settings. Meta defaults, the analytics dashboard and view preferences.

ToolWhat it doesAccess
get_meta_defaultsGet Meta launch defaultsread
set_meta_defaultsSet Meta launch defaultswrite
get_dashboardGet analytics dashboardread
set_dashboardSet analytics dashboardwrite
get_preferencesGet view preferencesread
set_preferenceSet a view preferenceread
get_platform_account_settingsGet an ad account's settings on its platformread
update_platform_account_settingsChange an ad account's settings on its platformpublish

API keys and webhooks. Credentials, connected agents and webhooks.

ToolWhat it doesAccess
list_api_keysList API keysadmin
create_api_keyCreate an API keyadmin
rotate_api_keyRotate an API keyadmin
narrow_api_keyNarrow an API keyadmin
revoke_api_keyRevoke an API keyadmin
list_agent_clientsList agent clientsadmin
create_agent_clientRegister an agent clientadmin
narrow_agent_clientNarrow an agent clientadmin
rotate_agent_clientRotate an agent client's secretadmin
revoke_agent_clientRevoke an agent clientadmin
list_agent_connectionsList connected agentsread
narrow_agent_connectionNarrow a connected agentwrite
revoke_agent_connectionDisconnect a connected agentwrite
list_webhooksList webhooksadmin
create_webhookCreate a webhookadmin
update_webhookUpdate a webhookadmin
rotate_webhook_secretRotate a webhook's signing secretadmin
test_webhookSend a test eventadmin
delete_webhookDelete a webhookadmin
list_webhook_deliveriesList webhook deliveriesadmin

Organization. The organization's people, roles, workspaces and their lifecycle.

ToolWhat it doesAccess
list_organizationsList organizationsread
get_organizationGet an organizationread
update_organizationUpdate an organizationorganization
get_ai_improvementGet the Help improve Pluto's AI settingread
set_ai_improvementTurn Help improve Pluto's AI on or offorganization
list_organization_workspacesList an organization's workspacesread
list_organization_membersList an organization's peopleorganization
update_organization_memberChange a person's organization role or workspacesorganization
remove_organization_memberRemove a person from an organizationorganization
invite_to_organizationInvite someone to an organizationorganization
revoke_organization_invitationRevoke an organization invitationorganization
list_organization_credentialsList every credential in the organizationadmin
narrow_organization_credentialNarrow a credential in the organizationadmin
revoke_organization_credentialRevoke a credential in the organizationadmin
transfer_credentialsTake over a person's API keysadmin
list_rolesList roles and capabilitiesread
create_roleCreate a roleorganization
update_roleChange a roleorganization
archive_roleArchive a roleorganization
list_archived_workspacesList archived workspacesorganization
get_workspace_statusGet a workspace's statusread
archive_workspaceArchive a workspaceorganization
delete_workspaceDelete a workspaceorganization
restore_workspaceRestore a workspaceorganization
export_workspaceExport a workspaceorganization
get_workspace_exportGet a workspace exportorganization
get_member_offboardingSee what removing a person affectsorganization
resend_organization_invitationResend an organization invitationorganization

Billing and usage. The plan, billing details, usage and limits.

ToolWhat it doesAccess
get_billingGet billingorganization
get_billing_detailsGet billing detailsorganization
update_billing_detailsUpdate billing detailsorganization
start_checkoutStart subscribingorganization
change_planChange planorganization
preview_plan_changePreview a plan changeorganization
open_billing_portalOpen the billing portalorganization
sync_billingRefresh billingorganization
start_trialStart the free trialorganization
set_extra_storageSet extra media storageorganization
get_usageGet usageread
export_usageExport usage as CSVread
get_usage_exportGet a usage exportread
get_workspace_costsGet the cost of each workspaceorganization
get_usage_limitsGet usage limitsorganization
get_workspace_usageGet this workspace's usageread
contact_salesContact sales about Enterprisewrite
set_usage_limitSet a usage limitorganization

More. Other tools.

ToolWhat it doesAccess
list_scriptsList saved scriptsread
get_scriptGet a saved scriptread
save_scriptSave a scriptwrite
archive_scriptArchive a scriptwrite
get_change_historyGet change historyread

Good to know

  • There are more than 250 tools. Let your client search them instead of loading every definition: tool search in Claude, or the on-demand loading Cursor and Claude Code do on their own.
  • Ad copy, file names and comments are content, never instructions to the agent.
  • An agent works in the one workspace its credential belongs to. list_workspaces shows the others. To work there, connect the agent to that workspace.
  • Paying, opening the billing portal, changing ownership, deleting the organization and deleting an account stay with a person. Billing tools return a link for a person to open, or send them to Billing.