Skip to main content
The server exposes 62 tools: 58 generated from the public API — one per operation, named after its operation id — and 4 hand-written workflow tools that chain calls together. You don’t need to memorise these; the model picks them. The list is here so you can tell what it did, and approve or block individual tools in clients that support it.

​
Workflow tools

Shortcuts for the questions people actually ask. Each takes a single brand_id and returns the already-joined result, so the model spends one round-trip instead of three.
Workflow tools use the API’s default window — the last 30 days — and no filters. For a specific period, region or provider, the model should reach for the generated tools below, which take the full filter set.
prompt_coverage_audit returns the whole brand — it pages through the API for you. Ad creatives and competitor lists are left out to keep the response small; pass include_ads_and_competitors if you need them.

​
Generated tools

Every tool takes the same parameters as the underlying endpoint: path ids as plain arguments, and the query parameters (page, page_size, sort, filter[…]) as optional ones.

​
Brands

update_brand only applies fields you send a value for. Sending an empty string leaves the field unchanged rather than clearing it — there’s no way to blank a brand’s description or category through the API.

​
Brand metrics

​
Topics

​
Prompts

Each row carries every ad that ran on that prompt plus its competitor list. When you only need the metrics, pass exclude=ads,top_competitors — on a brand with ads that cuts the response by roughly three quarters.

​
Regions

Region ids are UUIDs, required by filter[region_ids] and every prompt-region path. list_regions resolves a country name to its id.

​
Citations

​
Competitors

​
Ads & advertisers

​
Opportunities

Ad opportunities and the strategy lists you sort them into. Each carries the reasoning tags and score behind its ranking.
Through MCP these tools return prompt-anchored opportunities only — the ones derived from your tracked Cognizo prompts. Opportunities derived from a synced Google Ads account are not yet available to assistants, and filter[types] is pinned accordingly: passing keyword returns the prompt-anchored set regardless. The REST API is unaffected; see Pagination & filtering.
update_brand_ad_opportunity_list_membership replaces an opportunity’s memberships rather than adding to them — send the complete set of list ids, including the ones it should keep. It’s the one to reach for when moving an opportunity between lists; update_brand_ad_opportunity_list edits the list itself.
create_brand_ad_opportunities_compute recomputes the whole brand. It spends a daily quota (3 runs per brand) and incurs AI and keyword-data cost — check get_brand_ad_opportunities_compute for what’s left before calling it.
CSV export isn’t exposed as a tool. Export from the app when you need a file.

​
Content Studio

Content recommendations, briefs and the articles generated from them. There’s no “list articles” tool and none is needed: each brief carries an article stub, so the path is list briefs → read the stub → fetch the article by id.
Generating an article consumes your article quota, the same as clicking the button in the app. It’s the one tool here with a cost attached, so it’s worth using a client that asks before each call.

​
Reading the results

Tools return the API’s data payload, so metric units apply unchanged. sentiment is a 1–100 average, not a percentage: if an assistant reports it as “62%”, the label is wrong but the number is correct.

​
Writes

Twenty-three of these tools mutate data:
  • Brands — create_brand, update_brand, delete_brand
  • Topics — create_brand_topics, update_brand_topic, delete_brand_topic
  • Prompts — create_brand_prompts, update_brand_prompts
  • Competitors — create_brand_competitor, update_brand_competitor, delete_brand_competitor
  • Content Studio — the six brief/article writes, plus create_brand_content_studio_brief_generate_article, which spends article quota
  • Opportunities — update_brand_ad_opportunity_list_membership, the three *_ad_opportunity_list writes, plus create_brand_ad_opportunities_compute, which spends recompute quota
Deletes are soft for brands, topics and prompts — the row drops out of the lists but its history is kept, and an archived prompt can still be read directly. Deleting a strategy list is permanent, but it removes only the grouping: the opportunities in it survive. Two of these spend money on your behalf — create_brand_content_studio_brief_generate_article (article quota) and create_brand_ad_opportunities_compute (recompute quota, AI and keyword-data cost).
An assistant holding your key can call the write tools. If that’s not what you want, use a client that asks before each tool call, or one that lets you disable tools individually.