Skip to main content
Glama
Upload-Post

Upload-Post

Official

Manage ChatGPT Ads

manage_chatgpt_ads
Destructive

Create, update, pause or archive ChatGPT Ads campaigns, ad groups and ads in a connected ad account, with a preview-and-approval step before changes are sent.

Instructions

Create and change ChatGPT Ads campaigns, ad groups and ads in the ad account the user connected to Upload-Post. This spends the user's money: spend runs on their own OpenAI billing, and an ad serves only when its campaign, ad group and ad are all active and OpenAI approved the ad. Objects are created paused unless params.status is active. Fields go in params with the API's names. promote: campaign + ad group + ad in one call (title 3-50 characters, body up to 100, target_url, image_url or file_id; budget = lifetime spend limit in the account currency, minimum 1, or budget_micros; max_bid per billing event or max_bid_micros; bidding_type impressions|clicks|conversions, billing_event_type impression|click, start_time, end_time, locations, targeting, context_hints, name, ad_group_name, ad_name, status); if a later step fails the error lists in details.created what was already created. create_campaign (name, budget or budget_micros, status, description, start_time, end_time, bidding_type, conversion_event_setting_ids, locations, custom_audience_ids, excluded_custom_audience_ids, targeting); create_ad_group (campaign_id, name, max_bid or max_bid_micros unless bidding_config, billing_event_type, status, description, context_hints, product_set); create_ad (ad_group_id, name, and title, body, target_url, file_id, price at the top level or a full creative object; type chat_card|product_ad_template; it serves after OpenAI's review); upload_creative (image_url, OpenAI fetches it; returns the file_id an ad uses); update_campaign / update_ad_group / update_ad (the id plus only the fields to change; an edited ad goes back into review); activate_* starts spending, pause_* stops it, archive_* is irreversible (campaign_id, ad_group_id or ad_id). Connecting or disconnecting the ad account is done in the Upload-Post dashboard. Read results with chatgpt_ads_read. Needs the user's approval: call it first without confirmed to get a preview (nothing is sent), show the user exactly what will go out, and repeat the call with confirmed: true only after they approve.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "action": "<action>"
    },
    "name": "manage_chatgpt_ads"
  }
}

Input Schema

NameRequiredDescriptionDefault
ad_idNoThe ad (ad, ad_insights, preview_ad, update/activate/pause/archive_ad).
actionYes
paramsNoOther API fields for this action, with the REST API's own names (snake_case) and shapes. See the description for each action's fields. Leave out what you do not want to change.
confirmedNoSet to true only after the user has seen the preview this tool returns without it and approved it. Without it nothing is sent.
ad_group_idNoThe ad group (ad_group, ad_group_insights, update/activate/pause/archive_ad_group; ads lists the ads of this group; create_ad puts the ad in it).
campaign_idNoThe campaign (campaign, campaign_insights, update/activate/pause/archive_campaign; ad_groups lists the groups of this campaign; create_ad_group puts the group in it).
ad_account_idNoConnected ad account to use (`adacct_...`, from chatgpt_ads_read accounts). Without it and without profile_username: the default account.
profile_usernameNoUse the ad account linked to this Upload-Post profile, and only that one.

Output Schema

NameRequiredDescriptionDefault
resultNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv0.26.5

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well beyond the annotations (readOnlyHint=false, destructiveHint=true, openWorldHint=true): it discloses that real money is spent on the user's OpenAI billing, that an ad serves only when campaign, ad group and ad are all active and OpenAI has approved it, that objects are created paused unless status is active, that archive_* is irreversible, and that an edited ad re-enters review.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Purpose and money risk are front-loaded, which is good, but the core is one enormous semicolon-chained parenthetical listing every action's fields, making it hard to scan and forcing re-reads. The approval paragraph duplicates the `confirmed` schema description rather than delegating to it.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a high-blast-radius, 17-action mutation tool, the definition covers everything an agent needs: per-action fields, the paused-by-default behavior, partial-failure recovery via details.created, the irreversibility of archive_*, and the preview/confirm gate. Return values can be omitted since an output schema exists.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Despite 88% schema coverage, the description adds substantial meaning the schema cannot: that action-specific fields belong in `params` under the REST API's own names, enumerated field lists per action, character limits (title 3-50, body up to 100), budget as a lifetime spend limit with a minimum of 1, and the budget/budget_micros and max_bid/max_bid_micros equivalences.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb set and resource set ('Create and change ChatGPT Ads campaigns, ad groups and ads') plus the account scope ('the ad account the user connected to Upload-Post'), then names the companion read tool explicitly ('Read results with chatgpt_ads_read'), so an agent can separate it from the sibling without inspecting either schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives concrete routing rules: use promote for a one-shot campaign+group+ad, per-action calls otherwise, chatgpt_ads_read for results, and the Upload-Post dashboard for connecting/disconnecting the account. It also states the mandatory approval protocol (call without `confirmed` for a preview, show the user, then repeat with `confirmed: true`).

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools