> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kime.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP use cases

> Real workflows you can run with the KIME MCP server, from visibility reports and prompt library cleanup to ChatGPT ad campaigns, plus every tool and its scope.

The KIME MCP server exposes 50 tools, all prefixed `kime_*`, that an MCP-compatible AI assistant (Claude, ChatGPT, Cursor, or a custom agent you've built) can call on your behalf. You ask in plain language, and your assistant reads your workspace, connects the data, explains what it finds and, when you ask, takes action: adding and organizing prompts, adding competitors, managing actions, or drafting ad campaigns.

Every tool works on one workspace. Your assistant resolves the workspace, brands, categories, and tags once with `kime_list_workspaces`, `kime_list_brands`, `kime_list_categories`, and `kime_list_tags`, and reuses those IDs for the rest of the conversation.

> [!NOTE]
>
> **What you can change:** add, archive, and restore prompts; create categories and tags and assign them to prompts in bulk; add competitors; create and update actions and comment on them; and generate ad campaign proposals. You can't edit a prompt's text or permanently delete anything. To make changes you must be a member of the organization and select it when you connect.

## Why MCP instead of the dashboard

The dashboard is built for looking things up. MCP is built for asking questions and getting things done without switching tools. The same underlying data is available either way. MCP puts a conversational layer, and the ability to take action, in front of it. That matters most when:

* The person who needs the answer isn't the person who lives in the KIME dashboard day-to-day (an executive, a client, a teammate on another team).
* The question requires connecting several pieces of data together ("why" questions, not just "what" questions).
* You want the assistant to do something with the answer, such as file an action, retag 300 prompts, or draft a campaign, not just report a number.
* You're already working inside another tool (an IDE, a chat app, an internal agent) and don't want to leave it.

## Brand visibility and competitive benchmarking

**Tools:** `kime_list_brands`, `kime_get_performance`, `kime_get_visibility`, `kime_get_share_of_voice`

**You ask:** *"How's our AI visibility trending, and how do we compare to \[Competitor]?"* or *"Give me a one-line performance summary I can put in a status update."*

Your assistant resolves your brand and the named competitor with `kime_list_brands`, then combines `kime_get_visibility` (a time series of how often each brand shows up in AI answers), `kime_get_share_of_voice` (your share of brand mentions across the same prompts), and `kime_get_performance` (your visibility, position, sentiment, and share-of-voice scores) into a plain-language answer.

These tools accept a date range, an AI model, a category, tags, and a location, so you can go narrower in the same conversation: *"Now just for the enterprise category"* or *"Only ChatGPT, only Germany."* Your assistant re-runs the same calls with the added scope.

In clients that support MCP Apps, such as Claude, ChatGPT, and VS Code, `kime_get_visibility` and `kime_get_share_of_voice` render as an interactive KIME chart directly in the conversation.

## AI perception diagnostics: not just a score, but why

**Tools:** `kime_get_sentiment`, `kime_get_sentiment_breakdown`, `kime_get_brand_insights`, `kime_get_topic_keywords`, `kime_get_citation_analysis`

**You ask:** *"Why did our sentiment score move this period?"* or *"What's driving negative sentiment right now?"*

Your assistant chains several calls. `kime_get_sentiment` spots the trend. `kime_get_sentiment_breakdown`, filtered to negative excerpts, finds the actual sentences behind the drop: each excerpt is scored on eight factors (lexical tone, stance, comparative framing, supporting evidence, freshness, risk language, uncertainty, and salience) and linked to the source URLs it came from. `kime_get_topic_keywords` shows which recurring words and phrases drive positive or negative tone, for your brand or for any competitor.

Instead of a number that says "sentiment: moderate," you get something like: *"Two themes are pulling your score down. A handful of sites frame your pricing as steep relative to alternatives, and one recurring comparison page positions a competitor as more established. Here are the specific pages, if you want to follow up."*

`kime_get_brand_insights` returns an AI-generated narrative summary of how AI models perceive your brand, which works well as the opening of an executive brief.

## Read and fact-check the AI answers

**Tools:** `kime_list_prompt_chats`, `kime_get_chat`, `kime_get_prompt`

**You ask:** *"What exactly did ChatGPT and Gemini say about our pricing last week?"* or *"Check the AI answers for our product prompts against this spec sheet and list every factual error."*

Every metric in KIME comes from real AI answers, and your assistant can read them. `kime_list_prompt_chats` returns recent answers with their prompt, model, and run status, filtered by prompt, brand, AI model, or date range. `kime_get_chat` returns one full answer. `kime_get_prompt` adds the industry ranking for a single prompt, so you can ask who currently leads a query and why.

This turns "our visibility dropped" into "here is what the answer says now, and here is what it said before." Combined with your own documents, your assistant can find outdated or wrong claims about your products and suggest what to publish to correct them.

## Citation and source research on demand

**Tools:** `kime_list_source_domains`, `kime_get_citation_analysis`

**You ask:** *"Which sites are AI models citing when our category comes up?"* or *"Which of those are negative toward us?"*

`kime_list_source_domains` returns the domains feeding AI answers on your tracked prompts, with how often each is used, and can be scoped by category, tags, or location. `kime_get_citation_analysis` drills into the sources cited alongside your brand, with citation counts, average sentiment, and a sentiment-band distribution per source.

A PR or content lead gets back a ranked outreach list (the sites most worth pitching for a mention or a correction) grounded in actual citation frequency, and can immediately follow up with *"which of those are sentiment-negative toward us?"* to prioritize damage control.

## Build and organize your prompt library

**Tools:** `kime_bulk_create_prompts`, `kime_get_prompt_run_status`, `kime_list_prompts`, `kime_bulk_create_categories`, `kime_bulk_create_tags`, `kime_bulk_update_prompt_categories`, `kime_bulk_update_prompt_tags`, `kime_archive_prompts`, `kime_unarchive_prompts`

**You ask:** *"We're launching in Denmark. Set up tracking for our branded and comparison queries there."* or *"Clean up our library: group every prompt into a category and tag the ones about integrations."*

Your assistant turns a plain-language request into a batch of up to 500 prompts, each with the right country and, optionally, a category and tags, and submits them in one `kime_bulk_create_prompts` call. It then polls `kime_get_prompt_run_status` until every model has answered and tells you when results are in.

Organizing works the same way. `kime_bulk_create_categories` and `kime_bulk_create_tags` create up to 100 at a time and reuse any that already exist, so your assistant can safely send the full list it needs. `kime_bulk_update_prompt_categories` and `kime_bulk_update_prompt_tags` then assign them to up to 500 prompts per call, without re-running those prompts. To retire prompts, `kime_archive_prompts` stops them from running and frees up your prompt limit; `kime_unarchive_prompts` brings them back.

> [!WARNING]
>
> Creating prompts is a real write action, not a preview: each new prompt starts live runs on every AI model enabled in your workspace and counts toward your prompt limit. A well-built assistant reads the proposed list back to you and gets confirmation first, especially for large batches.

## Grow your competitor set

**Tool:** `kime_bulk_create_competitors`

**You ask:** *"Add the 15 brands that keep showing up next to us in AI answers as competitors."*

Your assistant can read the AI answers and rankings to find who appears alongside you, then add up to 500 competitors in one call with a name and domain each. Domains are normalized, and duplicates of a competitor you already track are skipped instead of failing the request.

## Instant technical GEO audits

**Tool:** `kime_get_ai_website_health`

**You ask:** *"Can AI crawlers actually read our site?"* or *"What should we fix first to be easier for AI to cite?"*

This returns your workspace's [AI Website Health](https://docs.kime.ai/features/ai-website-health.md) results: overall and category scores, domain and page checks (including whether `robots.txt` lets crawlers such as GPTBot, ClaudeBot, and PerplexityBot in), and the pages that were crawled. Your assistant turns the failed checks into a prioritized fix list, and can file each fix as an action in the same conversation. Run a new check from the AI Website Health page in KIME.

## Conversational Actions Center triage

**Tools:** `kime_list_tasks`, `kime_get_task`, `kime_bulk_get_tasks`, `kime_create_task`, `kime_update_task`, `kime_add_task_comment`, `kime_list_task_comments`

**You ask:** *"What should we work on this week?"* then keep going conversationally.

KIME's Actions Center turns gap analysis into a prioritized backlog automatically: content gaps, comparison-page updates, "get mentioned here" outreach opportunities, and validation and trust-signal suggestions, each with a full brief. Your assistant reads and explains this list in plain language instead of you scanning a table, then acts on it in the same conversation:

* *"Create a task to fix that gap"* uses `kime_create_task`.
* *"Mark it in progress, raise it to high priority, and note the angle"* uses `kime_update_task` plus `kime_add_task_comment`.
* *"Show me everything high-priority and still unassigned"* reads `kime_list_tasks` and filters the result.
* *"Pull up the three tasks I flagged yesterday"* uses `kime_bulk_get_tasks`.

The audit trail (status changes and comments) lives in KIME, so your team sees the same backlog in the dashboard and in any other tool connected to it.

## Review AI-generated content inline in chat

**Tools:** `kime_list_content_tasks`, `kime_get_content_task_output`

**You ask:** *"Is that article ready? Show me the draft."*

Actions can come with a linked [deliverable](https://docs.kime.ai/features/deliverables.md): an article, email, or social post that KIME's agents write for you. `kime_list_content_tasks` lists them with their status, and `kime_get_content_task_output` returns the finished output. For articles that's the full article with its SEO, social, and structured-data fields; for emails it's the body plus fields such as subject and recipient.

Articles export as JSON, Markdown, HTML, or WordPress's WXR import format, so once you approve a draft in chat, the same assistant (or another MCP server you've connected, such as your CMS) can take it straight to publishing.

## Competitive ads intelligence

**Tools:** `kime_get_ads_summary`, `kime_get_ads_trend`, `kime_get_ads_breakdown`, `kime_get_prompt_ads_runs`, `kime_list_ads_advertisers`, `kime_get_ads_advertiser`, `kime_list_ads_creatives`, `kime_get_ads_creative`

**You ask:** *"Is anyone running ChatGPT ads on our prompts, and who?"* or *"Where does \[advertiser] show up, and what do their ads say?"*

KIME tracks the sponsored ads ChatGPT shows on your prompts, separately from organic mentions (see [Ads Library](https://docs.kime.ai/features/ads-library.md)). Your assistant can:

* Report coverage with `kime_get_ads_summary`: how many prompts served ads, ad units, advertisers, and creatives, compared with the previous period.
* Show the weekly trend with `kime_get_ads_trend`.
* Find where paid competition is concentrated with `kime_get_ads_breakdown`, grouped by category, tag, prompt, or country, optionally for one advertiser.
* Show the week-by-week ad history of one prompt with `kime_get_prompt_ads_runs`, including position, creative, advertiser, and click URL.
* Rank advertisers and drill into one with `kime_list_ads_advertisers` and `kime_get_ads_advertiser`, including a per-country breakdown.
* Browse ad copy and imagery with `kime_list_ads_creatives` and `kime_get_ads_creative`.

That turns "is this becoming a paid channel we need to think about?" into a quick conversational check instead of a research project.

## From opportunity to ChatGPT ad campaign

**Tools:** `kime_get_campaign_builder_opportunity`, `kime_generate_campaign_proposals`, `kime_get_campaign_generation_status`, `kime_list_campaign_proposals`, `kime_get_campaign_proposal`, `kime_get_ads_manager_results`

**You ask:** *"Where should we buy ChatGPT ads to catch up with competitors? Draft campaigns for the best prompts."*

Your assistant can run the whole [Campaign Builder](https://docs.kime.ai/features/campaign-builder.md) flow in one conversation:

**1. Find where to play**

`kime_get_campaign_builder_opportunity` compares your organic strength with your competitors' on every prompt and puts each one in a target area: **defend** (you're strong, they're weak), **catch up** (they're strong, you're weak), **compete** (both strong), or **open** (nobody is strong).

**2. Confirm the shortlist**

Your assistant picks the prompts worth paying for, for example the biggest catch-up gaps, and confirms the selection with you, along with markets and ad language.

**3. Generate proposals**

`kime_generate_campaign_proposals` asks KIME to draft OpenAI Ads campaigns, ad groups, and ad copy from those prompts. `kime_get_campaign_generation_status` waits for the job to finish, which usually takes 15 to 60 seconds.

**4. Review the drafts**

`kime_list_campaign_proposals` and `kime_get_campaign_proposal` return each proposal's targeting, rationale, budget, ad copy, and image brief, so you can review and refine them in chat.

**5. Measure live results**

After you accept and publish campaigns in [Ads Manager](https://docs.kime.ai/features/ads-manager.md), `kime_get_ads_manager_results` reports impressions, clicks, spend, CTR, CPC, and CPM per campaign from your connected OpenAI Ads account.

> [!NOTE]
>
> Generating proposals never publishes anything or spends money. Proposals land on the Ads Manager **Proposed** tab, and you accept and publish them in KIME.

## Using it inside your coding tools, not just chat

Because this is MCP, it works anywhere an MCP client does: an IDE-integrated agent, a custom internal tool, or any assistant you've wired it into. A developer working on the marketing site can ask the same assistant, without leaving their editor, *"which pages are we missing that competitors have?"* and get an actioned list, or have the assistant open the relevant action while already looking at the page in question. If you also connect other MCP servers (your CMS, your project tracker, your analytics stack), the same assistant can reason across all of them in one conversation, for example cross-referencing a KIME content gap against your CMS's draft queue before creating a duplicate action.

## Notes for building on the MCP tools

* **Tool naming:** everything is prefixed `kime_*`. Discovery tools (`kime_list_workspaces`, `kime_list_brands`, `kime_list_categories`, `kime_list_tags`) return the IDs the other tools take.
* **Scoping filters:** most analytics tools accept `start_date`, `end_date`, `llm_model`, `category_id`, `tag_ids`, and `location`, so a narrower question doesn't need a different tool, only different arguments.
* **Pagination:** list tools return `items`, `totalCount`, `limit`, and `offset`. To get everything, page by incrementing `offset` by `limit` until it reaches `totalCount`.
* **Large responses:** a few tools, such as `kime_get_ai_website_health`, `kime_get_sentiment_breakdown`, `kime_list_prompt_chats`, and `kime_get_citation_analysis`, can return large payloads on a heavily tracked workspace. Narrow the date range or filters, and summarize rather than paste raw JSON.
* **Annotations:** every tool is marked read-only or write in its MCP annotations, so clients can ask for confirmation before any write.
* **Read versus write:** `read:kime` alone covers every question in this guide. Grant `write:prompts`, `write:brands`, and `write:tasks` only to an assistant you trust to act on your behalf, because they make real, persisted changes.

## Tool index

| Tool | Scope | What it does |
| - | - | - |
| `kime_list_workspaces` | `read:kime` | List the workspaces you can access. |
| `kime_list_brands` | `read:kime` | List your brand and tracked competitors. |
| `kime_list_categories` | `read:kime` | List prompt categories. |
| `kime_list_tags` | `read:kime` | List prompt tags. |
| `kime_get_performance` | `read:kime` | Performance scores for your brand. |
| `kime_get_visibility` | `read:kime` | Visibility over time for each brand, with an interactive chart. |
| `kime_get_share_of_voice` | `read:kime` | Share of voice for each brand, with an interactive chart. |
| `kime_get_sentiment` | `read:kime` | Sentiment over time for each brand. |
| `kime_get_sentiment_breakdown` | `read:kime` | Excerpt-level sentiment with eight factors and sources. |
| `kime_get_brand_insights` | `read:kime` | AI-generated summary of how AI perceives your brand. |
| `kime_get_topic_keywords` | `read:kime` | Keywords and phrases driving sentiment. |
| `kime_get_citation_analysis` | `read:kime` | Sources cited about your brand, with sentiment. |
| `kime_list_source_domains` | `read:kime` | Domains cited in AI answers, with usage counts. |
| `kime_list_prompts` | `read:kime` | List prompts with their metrics. |
| `kime_get_prompt` | `read:kime` | One prompt with full metrics and industry ranking. |
| `kime_get_prompt_run_status` | `read:kime` | Run status per AI model for up to 100 prompts. |
| `kime_list_prompt_chats` | `read:kime` | AI answers for your prompts, with filters. |
| `kime_get_chat` | `read:kime` | One full AI answer. |
| `kime_bulk_create_prompts` | `write:prompts` | Add up to 500 prompts and start tracking them. |
| `kime_archive_prompts` | `write:prompts` | Archive up to 500 prompts. |
| `kime_unarchive_prompts` | `write:prompts` | Restore up to 500 archived prompts. |
| `kime_bulk_create_categories` | `write:prompts` | Create up to 100 categories. |
| `kime_bulk_create_tags` | `write:prompts` | Create up to 100 tags. |
| `kime_bulk_update_prompt_categories` | `write:prompts` | Set the category of up to 500 prompts. |
| `kime_bulk_update_prompt_tags` | `write:prompts` | Set the tags of up to 500 prompts. |
| `kime_bulk_create_competitors` | `write:brands` | Add up to 500 competitors. |
| `kime_list_tasks` | `read:kime` | List actions. |
| `kime_get_task` | `read:kime` | One action. |
| `kime_bulk_get_tasks` | `read:kime` | Up to 100 actions by ID. |
| `kime_list_task_comments` | `read:kime` | Comments on an action. |
| `kime_create_task` | `write:tasks` | Create an action. |
| `kime_update_task` | `write:tasks` | Update an action's status, priority, assignee, and more. |
| `kime_add_task_comment` | `write:tasks` | Comment on an action. |
| `kime_list_content_tasks` | `read:kime` | List generated articles, emails, and social posts. |
| `kime_get_content_task_output` | `read:kime` | Read or export generated content. |
| `kime_get_ai_website_health` | `read:kime` | AI Website Health scores and checks. |
| `kime_get_ads_summary` | `read:kime` | ChatGPT ads coverage for your prompts. |
| `kime_get_ads_trend` | `read:kime` | Weekly ads coverage. |
| `kime_get_ads_breakdown` | `read:kime` | Ads coverage by category, tag, prompt, or country. |
| `kime_get_prompt_ads_runs` | `read:kime` | Weekly ad history of one prompt. |
| `kime_list_ads_advertisers` | `read:kime` | Advertisers ranked by reach. |
| `kime_get_ads_advertiser` | `read:kime` | One advertiser, with trend and countries. |
| `kime_list_ads_creatives` | `read:kime` | Ad creatives gallery. |
| `kime_get_ads_creative` | `read:kime` | One ad creative. |
| `kime_get_campaign_builder_opportunity` | `read:kime` | Where to play: target areas per prompt. |
| `kime_generate_campaign_proposals` | `write:tasks` | Draft OpenAI Ads campaign proposals. |
| `kime_get_campaign_generation_status` | `read:kime` | Wait for proposal generation to finish. |
| `kime_list_campaign_proposals` | `read:kime` | List proposed campaigns. |
| `kime_get_campaign_proposal` | `read:kime` | One proposal with ad groups and ads. |
| `kime_get_ads_manager_results` | `read:kime` | Live results from your OpenAI Ads account. |

## When to reach for the API instead

MCP is the right layer when a person is in the loop, asking questions in natural language, or reviewing and approving before something happens. When you want something to run unattended (on a schedule, inside a pipeline, embedded in your own product), call the KIME API directly instead. See [API use cases](https://docs.kime.ai/api-use-cases.md) for that side of the picture, and [API access](https://docs.kime.ai/api-access.md) for what the REST API supports today.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.