# Connect AI assistants Source: https://docs.kime.ai/administration/integrations Give AI tools like ChatGPT and Claude secure access to your Kime data. Kime exposes a secure connection that lets AI assistants work with your data directly. Once connected, tools like Claude and ChatGPT can read your AI perception data, analyze your competitors, and help you act on your visibility — right inside the assistant you already use. You'll find it under **Integrations** in your organization settings. This integration is rolling out gradually. If you don't see **Integrations** in your organization yet, it isn't enabled for your account. ## Connection URL Every client connects to the same Kime endpoint. Your exact URL is shown on the **Integrations** page in your organization settings — always copy it from there so it's correct for your account and environment. ```text theme={null} https:// ``` ## Authentication Kime uses **OAuth** — there's nothing to paste or store. The first time an assistant uses a Kime tool, it opens a secure sign-in window where you log in and approve access. After that, your session persists across conversations. Access always respects your Kime account: an assistant can only see the organizations and workspaces you can, and acts within your existing permissions. ## Claude (Desktop & Web) In Claude, go to **Customize** (gear icon) → **Connectors**. Click **+**, choose **Add custom connector**, name it **Kime**, and paste the connection URL from the Integrations page. Click **Add**, then **Connect**. You'll be redirected to Kime to sign in. The first time you ask about your data, approve access when prompted. In some Claude workspaces, only an admin can add connectors. If you can't add one, ask your Claude workspace admin. ## ChatGPT Custom connectors in ChatGPT require **Developer mode** (available on paid plans; on Business and Enterprise workspaces an admin may need to enable it first). Go to **Settings** → **Apps** → **Advanced settings** and turn on **Developer mode**. Click **Create app**, name it **Kime**, paste the connection URL from the Integrations page, and choose **OAuth** as the authentication method. Save the app and use **Scan tools** to confirm the connection. In a new chat, open the **+** menu → **Developer mode** and enable your Kime connector. ## Other MCP clients Kime uses the standard **Model Context Protocol** over Streamable HTTP, so it works with any MCP-compatible tool — including Cursor, VS Code (GitHub Copilot), and Claude Code. Point your client at the Kime connection URL, choose **Streamable HTTP** as the transport, and sign in via OAuth when prompted. ```bash Claude Code theme={null} claude mcp add kime --transport http https:// ``` ```json VS Code (.vscode/mcp.json) theme={null} { "servers": { "kime": { "type": "http", "url": "https://" } } } ``` In **Cursor**, open **Settings → Tools & Integrations → MCP**, add a custom MCP server with the Kime URL, and select Streamable HTTP. You'll be prompted to sign in via Kime on first use. ## Verify your connection After connecting, ask your assistant something like: > "List my Kime workspaces." If it returns your workspaces, you're connected. From there you can ask it to summarize your visibility, compare competitors, or review your open actions. ## Troubleshooting Confirm you signed in with the correct Kime account, then remove and re-add the connector. Clearing cookies for the Kime sign-in window and reconnecting can also help. Make sure your Kime account has access to at least one workspace, and that you approved access for the right organization during sign-in. Check your internet connection and confirm the URL exactly matches the one on the Integrations page. If your client supports multiple transports, make sure it's set to Streamable HTTP. The integration is in gradual rollout. If it isn't visible in your organization settings, it isn't enabled for your account yet. # Roles & permissions Source: https://docs.kime.ai/administration/roles-and-permissions Control what each member of your team and your clients can see and do. Kime uses roles to control access so the right people can manage your workspace while others view results safely. ## Roles Full access — workspace settings, billing, user management, and data deletion. Can view all dashboards, manage prompts, and add competitors, but can't change workspace-level settings. ## Inviting people Invite teammates from your organization or workspace settings and assign a role based on what they need to do. Members are ideal for analysts and account managers who run day-to-day work; reserve Owner for those who manage configuration, billing, and access. ## Client (viewer) access On **Agency** accounts, you can invite a client to a workspace as a **viewer**. Viewers get read-only access to that workspace's dashboards so they can follow their metrics without changing the tracking setup. Manage client access under **Client Access** in Workspace settings. See [Setting up your first client](/agency/setting-up-your-first-client) for the full flow. | Capability | Owner | Member | Viewer | | ---------------------------- | ----- | ------ | ------ | | View dashboards | Yes | Yes | Yes | | Manage prompts & competitors | Yes | Yes | No | | Change workspace settings | Yes | No | No | | Billing & user management | Yes | No | No | Give new teammates the Member role by default and reserve Owner for the people who manage billing and access. # Creating pitch workspaces Source: https://docs.kime.ai/agency/creating-pitch-workspaces Use KIME to generate immediate competitive insights for prospective client pitches. Kime is a powerful pitch tool. Before a meeting, you can spin up a workspace for a prospect and instantly show them how their brand performs in AI answers — and where competitors are beating them. ## Audit a prospect Enter a prospect's root domain when creating a workspace and Kime generates prompt and competitor suggestions automatically. Within a short time you have a baseline of their visibility, placement, and sentiment — no manual research required. This gives you the evidence to show a prospect exactly where they're invisible compared to their top competitors, using the [Industry ranking](/dashboard/industry-ranking) table to highlight gaps in placement and share of voice. ## Build the narrative The most persuasive parts of a pitch are usually: * **Mentions leaderboard** and **Visibility ranking** — show how a rival is outperforming the prospect right now. * **Latest model answers** — show what AI models are actually saying about the prospect today. If it's outdated or unflattering, you've found a clear reason for them to act. Demonstrating that a competitor owns the majority of the conversation in their category makes the case for your services on its own. ## Convert the pitch into an account When a prospect signs, you don't start over. Refine the existing pitch workspace — approve the most relevant suggestions, add campaign-specific prompts — and the data you pitched with becomes the foundation for ongoing reporting. Keep a few pitch workspaces warm for target accounts so you can walk into any meeting with live AI visibility data ready to show. # Managing multiple accounts Source: https://docs.kime.ai/agency/managing-multiple-accounts Scale your agency operations by efficiently organizing and navigating multiple client workspaces. As your client roster grows, Kime keeps every brand organized and isolated so your team can manage them all from a single login. ## Switch between clients Use the **workspace switcher** at the top of the sidebar to jump between client dashboards without logging out. Each workspace is fully isolated — prompts, competitors, and history for one client never mix with another. That separation is essential when you serve competing brands in the same industry. ## Allocate your team Assign team members to the specific workspaces they manage. Inviting analysts and account managers as **members** lets them run prompt libraries and analyze gaps without changing workspace-level settings, which keeps your data consistent across the roster. See [Roles & permissions](/administration/roles-and-permissions) for details. ## Standardize to scale A consistent framework across clients is what makes a large portfolio manageable: * Use the same category names and conventions in every workspace. * Onboard new team members faster because every account looks familiar. * Compare how similar categories perform across clients to spot broad shifts in AI behavior. ## Build agency-wide insight Managing many accounts lets your agency develop its own sense of what "good" looks like in each sector. Comparing the performance of your strongest clients against new prospect audits gives you realistic growth projections — and a track record that proves your expertise in AI visibility. Reuse a standard category template across clients so cross-account comparisons are always apples-to-apples. # Setting up your first client Source: https://docs.kime.ai/agency/setting-up-your-first-client Learn how to configure a dedicated workspace for client brands and manage team access. With an **Agency** account, you manage multiple client brands from one place. Each client lives in its own **workspace**, keeping prompts, competitors, and results fully isolated from your other clients. ## Create a client workspace Create a new workspace with the client's brand name and primary root domain. The domain anchors Kime's tracking and powers tailored prompt and competitor suggestions for that client's industry. Approve the proposed prompts and competitors that fit the client, just as you would for your own brand. Apply a consistent set of categories so every client workspace follows the same analytical framework. ## Manage client access Kime's roles let you collaborate with both your internal team and the client: * Invite your team as **members** to manage day-to-day optimization while you keep owner-level control. * Optionally invite client stakeholders as **viewers** so they can follow their metrics in read-only mode without changing the tracking setup. Client (viewer) access is managed per workspace under **Client Access** in Workspace settings. For the full role breakdown, see [Roles & permissions](/administration/roles-and-permissions). ## Standardize your reporting Consistency across clients makes your agency more efficient. Establish standard category names — such as "Product Category" or "Competitor Comparison" — and reuse them across every workspace. A shared framework makes it easy to compare performance and scale the strategies that work. Treat your first client workspace as a template. Once you've found a category structure that reports well, repeat it for every new client. # API access Source: https://docs.kime.ai/api-access API access to your data in Kime The Kime API gives you programmatic access to your AI visibility analytics. Use it to integrate Kime data into your own dashboards, BI tools, or internal systems. ### **What you can access** The API provides read-only access to your workspace data: | **Endpoint** | **Description** | | :------------- | :------------------------------------------------------------------- | | **Workspaces** | List workspaces in your organization | | **Brands** | List tracked brands in a workspace | | **Prompts** | List active prompts (queries) being monitored | | **Categories** | List prompt categories | | **Tags** | List prompt tags | | **Metrics** | Brand visibility, placement, sentiment, share of voice, and rankings | | **Domains** | Source domains cited by AI answer engines, with citation frequency | | **Sources** | Individual URLs cited, including which brands are mentioned | All analytics endpoints support filtering by date range, answer engine, brand, prompt, category, and tags. ### **Supported answer engines** Filter by any combination of: `chatgpt` · `perplexity` · `claude` · `gemini` · `google-ai-overview` · `google-ai-mode` ### **Authentication** The API uses **API keys** passed in the `x-api-key` header. Each key is scoped to your organization and can be restricted to specific resources. ### **How to get an API key** 1. Sign in to the [**Kime dashboard**](https://app.kime.ai/). 2. Go to **Settings → API Keys**. 3. Click **Create API Key** and give it a name. 4. Copy the key — it's shown only once. Your organization admin controls which API scopes are available. Keys can access only the resources their scopes permit (e.g., a key with `metrics` scope cannot access `sources`). ### **Making requests** All requests go to: [https://api.kime.ai/v1/](https://api.kime.ai/v1/)... Include your API key in the header: `x-api-key: your_api_key_here` Example — list your workspaces: `curl -H "x-api-key: your_api_key_here" \` [https://api.kime.ai/v1/workspaces](https://api.kime.ai/v1/workspaces) ### **Pagination** List endpoints return paginated results. Use `page_size` (max 100) and the `cursor` from the response to fetch additional pages. ### **Date filtering** Analytics endpoints default to the **last 7 days**. Override with `start_date` and `end_date` query parameters (format: `YYYY-MM-DD`). ### **Error handling** Errors return [**RFC 9457**](https://www.rfc-editor.org/rfc/rfc9457) Problem Details format with a `type`, `title`, `status`, and `detail` field. Common codes: | **Status** | **Meaning** | | :--------- | :------------------------- | | 401 | Invalid or missing API key | | 403 | Key lacks required scope | | 404 | Resource not found | | 422 | Invalid filter value | ### **Full API reference** Explore all endpoints, parameters, and response schemas in the interactive documentation: [**Open API Documentation →**](https://api.kime.ai/external/docs) ### **Need help?** Contact us at [team@kime.ai](mailto:team@kime.ai) # AI master overview Source: https://docs.kime.ai/dashboard/ai-overview The four headline scores that summarize your brand's health in AI answers. The AI master overview is the control center of your dashboard. It rolls up data from all of your tracked prompts and models into four headline scores that tell you, at a glance, how your brand is doing in AI answers. ## The four scores A single composite index of overall brand health, combining how often, how prominently, and how positively you're mentioned. How much of your tracked prompt library mentions your brand. How prominently you're positioned within AI answers, on average. How positive, neutral, or negative the tone is when AI mentions you. For a deeper definition of each, see the [Metrics](/metrics/visibility) section. ## Read them together The scores are most useful as a group: * **High visibility, weak placement** — AI knows your brand but isn't recommending it first. Focus on becoming a top recommendation. * **High visibility, low sentiment** — you're being talked about, but not favorably. Look at what's driving the negative tone in [AI Perception](/features/ai-perception). * **Rising performance score** — your overall health is improving; use the date filter to tie it back to a campaign or content push. ## Use it for reporting The master overview is the cleanest way to report progress to stakeholders. Adjust the date range to show how your scores moved after a launch or optimization effort, and use it as the consistent benchmark for your brand's growth in AI. # Dashboard overview Source: https://docs.kime.ai/dashboard/dashboard-overview A tour of the Kime workspace and the modules that show how your brand performs in AI answers. The dashboard is your home base in Kime. It brings your most important AI visibility signals into one view so you can move from a high-level health check to the exact text an AI generated about your brand in just a few clicks. ## Layout Kime is organized around a sidebar and a central workspace: * **Sidebar** — switch between areas like Prompts, Sources, Competitors, Competitor Analysis, AI Perception, Actions, and the Agent. * **Workspace** — modular cards that range from headline scores down to individual model answers and the sources behind them. ## Core modules Your headline health scores: performance, visibility, placement, and sentiment. How your mention volume stacks up against competitors. Visibility trends over time for you and your competitive set. A side-by-side table of every brand across your core metrics. The dashboard also surfaces the **latest model answers** and the **sources** behind them, so you can always trace a score back to the real AI responses that produced it. ## Filter what you see Controls at the top of the dashboard scope every module at once: | Filter | What it does | | --------------- | ---------------------------------------------------------------------------------- | | Date range | Switch between 7-, 14-, and 30-day views to see short-term spikes or longer trends | | Brand | Focus on your brand or a specific competitor | | Model | Isolate a single AI model, like ChatGPT or Perplexity | | Category & tags | Narrow to a product line or any tagged segment | Set the filters once at the top and the whole dashboard updates together — perfect for building a focused view before a stakeholder report. # Industry ranking Source: https://docs.kime.ai/dashboard/industry-ranking A side-by-side table comparing every brand across your core AI visibility metrics. The Industry ranking table puts your brand next to every competitor across all of your core metrics for the selected period. Where the [Visibility ranking](/dashboard/visibility-ranking) chart shows trends, this table gives you the exact numbers — making it the quickest way to see who leads on each dimension. ## The columns | Column | What it shows | | -------------- | -------------------------------------------------------- | | Visibility | Share of prompts where the brand was mentioned | | Placement | How prominently the brand appears in answers, on average | | Sentiment | The tone of the brand's mentions | | Share of Voice | The brand's slice of all mentions in the tracked set | Each metric also shows a trend indicator (green or red) versus the previous period, so you can tell who's gaining momentum and who's slipping. ## Analyze the table Sort by any column to find the leader on that dimension. Sorting by placement, for example, reveals which brand AI models favor as the top pick — regardless of how often it's mentioned. Use the table to find gaps in your strategy: * A competitor with **high sentiment but low visibility** is well-liked but under-mentioned. * If you have **high visibility but weaker placement** than rivals, focus on becoming the top recommendation rather than just appearing. Filter the dashboard by category before reading the table to find the specific niches where you lead — and the ones where a competitor is pulling ahead. # Mentions leaderboard Source: https://docs.kime.ai/dashboard/mentions-leaderboard Compare how often each brand in your competitive set is mentioned in AI answers. The Mentions leaderboard ranks every brand you track by how often it appears in AI answers. It's the fastest way to see who dominates the conversation in your category. ## What it shows For each brand, the leaderboard breaks mentions down by tone: * **Total mentions** — how often the brand appeared across your tracked prompts. * **Positive mentions** — mentions with a favorable tone. * **Negative mentions** — mentions with an unfavorable tone. Unless you apply filters at the top of the dashboard, the leaderboard aggregates across your selected models and markets. ## Trend indicators Each brand shows how its mention volume changed versus the previous period — green for growth, red for decline. These trends help you tell whether a competitor's recent marketing or content push is starting to influence what AI models say. ## Read the sentiment split More mentions isn't always better. A competitor with high volume but a large share of negative mentions may be getting attention for the wrong reasons. Watching the positive-to-negative split helps you spot reputation weaknesses — in their brand or your own — and decide where to focus. Pair the leaderboard with [Share of Voice](/metrics/share-of-voice) to see not just who's mentioned most, but what slice of the total conversation each brand owns. # Visibility ranking Source: https://docs.kime.ai/dashboard/visibility-ranking See how your visibility trends over time against your competitors. The Visibility ranking chart shows how your brand's presence in AI answers changes over time. Where the master overview is a snapshot, this chart is the trend line — essential for seeing the long-term impact of your work and catching sudden drops early. ## Compare multiple brands The chart plots your brand alongside the competitors you've selected. Hover over any point to see each brand's visibility on that day. Side by side, the lines reveal: * **Correlations** — whether the whole category moved together (often a sign of a model update). * **Outperformance** — periods where you gained while competitors stalled. * **Market shifts** — changes in which brands AI models favor over time. ## Interpret the movement Visibility is rarely a flat line; small day-to-day movement is normal as AI answers vary. Larger swings usually line up with real events: * **Spikes** often follow coverage on a high-authority site or publication that AI models draw from. * **Dips** can signal a competitor stepping up their content, or a shift in how a model surfaces brands. ## Choose your timeframe Use the 7-, 14-, and 30-day presets at the top of the dashboard. Longer windows are best for spotting seasonal patterns; the 7-day view is best for watching the immediate results of a recent campaign. # Data Studio Connector Source: https://docs.kime.ai/data-studio-connector Connect your Data Studio (Looker Studio) to Kime Bring your Kime AI visibility analytics directly into Looker Studio (formerly Google Data Studio) to build custom dashboards and reports. ### **What you can access** The connector pulls data from your Kime workspace via the Kime API: **Brand Metrics** * Visibility, placement, sentiment, and share of voice scores per brand * Period-over-period deltas for each metric * Brand ranking within your tracked competitive set **Source Domains** * Domains cited by AI answer engines (ChatGPT, Perplexity, Gemini, etc.) * Citation frequency and average citation position **Source URLs** * Individual URLs cited across answer engines * Which of your tracked brands are mentioned on each source * Citation counts and positions All data respects the date range you set in your Looker Studio report. ### **How to connect** 1. **Get a Kime API key** * In the Kime dashboard, go to **Settings → API Keys** and create a new key. * The key must have the `workspaces` scope plus the scope for the dataset you want (`metrics`, `domains`, or `sources`). * See [**API access**](https://docs.kime.ai/api-access) for details on creating keys. 2. **Add the connector in Looker Studio** * Click this link to add the Kime connector directly: [**Add Kime Connector to Looker Studio →**](https://lookerstudio.google.com/datasources/create?connectorId=AKfycbxjZpuqH7B_BVnVtNJBzgeq3sZ6k_zHYKRg2SD5KWS_cvIWq48goPIJZ1hAa0bTmqZFvQ) * When prompted, click **Authorize** to allow the connector to run. * Paste the Kime API key you created in step 1. 3. **Configure the data source** * Select your **Workspace** from the dropdown. * Choose a **Dataset**: Brand metrics, Source domains, or Source URLs. * Optionally filter by specific answer engines (e.g., `chatgpt`, `perplexity`, `claude`). * Click **Connect**. 4. **Build your report** * Use the date range control to filter by time period (defaults to last 7 days). * Add charts using dimensions like `Brand` or `Domain` and metrics like `Visibility` or `Used count` ### **Supported answer engines** Filter results by any combination of: `chatgpt`, `perplexity`, `claude`, `gemini`, `google-ai-overview`, `google-ai-mode` ### **Data refresh** Looker Studio fetches fresh data from Kime each time you view or refresh the report. There is no cached copy — you always see current analytics. ### **Need help?** * **API documentation**: [https://docs.kime.ai/api-access](https://docs.kime.ai/api-access) * **Support**: Contact us at [team@kime.ai](mailto:team@kime.ai) # AI Perception Source: https://docs.kime.ai/fearures/ai-perception **TL;DR:** KIME scores every AI response about your brand across 8 measurable dimensions: Language Tone, Competitive Position, Brand Focus, Endorsements, Source Credibility, Information Recency, Risk Factors, and Confidence Level. Each dimension shifts your score differently. Knowing which one is dragging you down is the difference between guessing and fixing. Most brand tracking tools tell you that AI sentiment changed. They do not tell you why. You get a score. It goes up, it goes down. And you are left guessing whether it was a new competitor entering the conversation, an old controversy resurfacing, or AI citing a three-year-old source that no longer reflects your product. **AI Perception** is KIME's solution to that problem, and it is available now. It gives you a structured, dimension-by-dimension breakdown of what AI outputs about your brand actually contain, scored using a deterministic algorithm and traceable to the real AI text behind every data point. Image 1 ## Why is not positive, neutral, negative enough? A three-bucket sentiment summary tells you the outcome, not the cause. It is like a doctor saying your bloodwork looks mostly fine without showing the actual values. AI perception is a composite signal, not a single thing. A brand can score well on Language Tone while scoring poorly on Competitive Position because every recommendation list places them third or fourth. The aggregate reads as neutral. The reality is specific, and fixable. Consider two brands in the same category. Brand A gets positive language but is always mentioned as an afterthought after two competitors. Brand B gets slightly uncertain language but is consistently the first recommendation. Standard sentiment says Brand A is positive. KIME shows Brand A has a Competitive Position problem that no amount of good press will fix on its own. To give you something actionable, you need to know which component is moving and in which direction. That is what AI Perception scores. Image 2 ## What are the 8 factors that drive your AI Perception score? When KIME analyzes an AI response about your brand, it applies a deterministic, rule-based scoring algorithm to the raw text. No black box, no vague AI sentiment. The algorithm measures 8 distinct dimensions of every response. Together they do not just tell you what AI is saying about you. They tell you why you scored the way you did, and what you can actually do about it. 1. **Language Tone** What it measures: the actual words AI chooses when describing your brand. Why it matters: it is the clearest signal of how AI has learned to speak about you. 1. **Competitive Position** What it measures: where AI places your brand in ranked lists, comparisons, and recommendations. Why it matters: most AI discovery happens through comparative queries. 1. **Brand Focus** What it measures: how much of the response is actually about you. Why it matters: a passing mention contributes far less than being the main subject of the answer. 1. **Endorsements** What it measures: whether AI explicitly recommends your brand or issues a caution. Why it matters: being an option is not the same as being the one to use. 1. **Source Credibility** What it measures: the authority level of the sources AI cites. Why it matters: a single mention in a tier-one publication can matter more than hundreds of low-authority mentions. 1. **Information Recency** What it measures: how current the information behind the AI answer actually is. Why it matters: AI can confidently describe a version of your brand that no longer exists. 1. **Risk Factors** What it measures: whether AI mentions controversies, legal issues, recalls, breaches, or other, product issues, product concerns. Why it matters: even small or outdated risks can insert doubt into otherwise positive answers. 1. **Confidence Level** What it measures: how much uncertain language language AI uses when talking about your brand. Why it matters: tentative language transfers uncertainty to the reader. ## How do you read your AI Perception score breakdown? Every brand's AI Perception score is a combination of these 8 factors, but they do not all contribute equally and they do not all move in the same direction at the same time. The KIME dashboard surfaces this as a ranked breakdown, so you can see at a glance what is helping and what is hurting. The strategic insight is that you do not need to optimize all 8 factors at once. The right approach is to identify the one or two factors with the largest negative influence on your current score, understand why they are underperforming, and address those first. A brand with strong Language Tone and Endorsement scores but a weak Source Credibility score has a different problem, and a different fix, than a brand with strong Source Credibility but poor Competitive Position. The score breakdown is designed to make that distinction obvious. ## Why does the same brand score differently across AI models? One pattern that consistently surprises brands is how much the score can vary across models. The same brand can look meaningfully different across models in the same time period. The practical takeaway is not that one number is right and another is wrong. It is that users encounter different answer environments across models, and your brand can be framed differently depending on where they ask. That is why AI perception should not be treated as a single number. Cross-model variance is valuable because it shows where your brand is strongest, where it is weakest, and where monitoring should be prioritized first. ## What should you do when a specific factor is pulling your score down? AI Perception is diagnostic by design. The score tells you where you are. The factor breakdown tells you why. But the key is that KIME does not stop at the label: it shows you the exact AI excerpts that created each signal, along with the sources the model cited in that answer. So when a factor is dragging you down, you are not guessing. You can see the wording, ranking, warning, or outdated claim itself, and the source behind it. 1. **Low Language Tone** - Audit the content AI is retrieving about you. Look for high-traffic negative reviews or critical editorial coverage. Build positive coverage from authoritative sources using specific, strong language. 2. **Low Competitive Position** - You are not appearing early enough in AI comparison outputs. Invest in shortlist and best-of editorial coverage. Improve third-party review scores on platforms AI uses as sources. 3. **Low Brand Focus** - AI does not have enough to say about you specifically. Create deeper, topic-specific content and build coverage where you are the subject, not a supporting mention. 4. **Low Endorsements** - AI mentions you but doesn't actively recommend you. Create case studies and get them picked up by authoritative third-party sites. Pitch "best for X" editorial coverage in recognized outlets in your category. 5. **Low Source Credibility** - Your brand is mentioned most by low-authority sources. Prioritize earned media in tier-one publications and recognized trade outlets. Get onto established review platforms. 6. **Low Information Recency** - AI is drawing on old information. Publish and distribute updated content, and actively pitch recent milestones and product updates. 7. **High Risk Factors** - A past issue is still surfacing in AI responses. Address it with authoritative, accurate content and build enough positive signal to reduce the proportional weight of risk mentions. 8. **Low Confidence Level** - Claims about your brand are not well corroborated. Get the same positive facts referenced across multiple independent, high-authority sources. The goal is to move from AI has an opinion about us and we cannot see it to AI has an opinion about us, we can see the exact text and cited sources behind it, and we know precisely what to do about it. ## Frequently Asked Questions **What is an AI perception score?** An AI perception score is a composite measure of what AI language models actually say about your brand in their outputs. KIME calculates it by applying a deterministic, rule-based algorithm to real AI responses and measuring 8 dimensions of each response: Language Tone, Competitive Position, Brand Focus, Endorsements, Source Credibility, Information Recency, Risk Factors, and Confidence Level. **Why do AI perception scores vary across models?** A brand that scores well in one model may score poorly in another because the answers users see can vary meaningfully across models. **How is AI perception different from traditional brand sentiment?** Traditional brand sentiment tracks how humans describe your brand in reviews, social media, and press. AI perception tracks how AI models describe your brand in their outputs, which is a different signal driven by different sources. **How often does an AI perception score change?** Scores can shift within days if a meaningful new piece of content, positive or negative, enters the sources AI retrieves from. KIME tracks your score over time so you can correlate changes with specific content and coverage events. **What is the fastest way to improve a low AI perception score?** The fastest improvement usually comes from addressing the single factor with the largest negative impact on your current score. Source Credibility and Information Recency tend to respond fastest to targeted action. ## How do you get started with AI Perception tracking? AI Perception is live in KIME today. Here is how to run your first analysis: 1. **Connect your brand** - Add your brand name and the competitor set you want to track. KIME starts pulling real AI responses immediately. 2. **Set your tracked queries** - Choose the prompts AI users are most likely to ask when evaluating your category. KIME scores your brand across all of them. 3. **Open your Perception dashboard** - Your factor breakdown appears automatically, with each dimension ranked by impact and the AI excerpts behind every score available behind a click. From there, sort by the factor with the largest drag, read the underlying AI text, and you know exactly what to address and why. Try KIME's AI Perception for free Talk to us # Actions Source: https://docs.kime.ai/features/actions Turn your AI visibility data into a prioritized, trackable plan of work. Actions is where Kime turns insight into execution. Instead of only measuring how your brand performs in AI answers, Actions gives you specific, prioritized recommendations for improving your visibility — organized on a board your whole team can work from. Kime generates fresh actions regularly, so you always have a current set of high-impact opportunities to work through. ## The board Actions is a Kanban-style board. Move a card across columns as the work progresses: Pending actions generated by Kime, prioritized by their potential impact on your visibility. Start with the highest-impact items. Assign an owner and move the card here to signal the work is underway. Once the requirements are met, move the card here to validate the work before it goes live. Completed actions are archived so you keep a record of your optimization journey and can track velocity over time. ## Inside an action Click any card to open a detailed brief that explains both the *why* and the *how*: * **Why it works** — the rationale for how the action improves your presence in AI answers. * **Must-haves** — the requirements to complete the action to Kime's standard. * **Could-haves** — optional enhancements that push performance further. * **Examples** — references to strong executions you can model your work on. ## Manage and collaborate Each card carries metadata you can adjust as your strategy evolves: | Field | Description | | -------- | ----------------------------------------------------- | | Assignee | The team member responsible | | Category | The type of work (e.g. editorial, content, technical) | | Priority | Urgency, from urgent to low | | Created | When Kime identified the opportunity | Suggested actions come with **Accept** and **Reject** controls so you can quickly curate your board. Use the filters at the top to focus by assignee, priority, or favorites, and the comments on each card to share drafts and feedback with your team. Want a recommendation explained or expanded in plain language? Ask the [Agent](/features/agent) — it can talk through your data and the actions behind it. # Agent Source: https://docs.kime.ai/features/agent Chat with Kime's AI agent to explore your data and get answers in plain language. The Agent is your conversational way into Kime. Instead of clicking through dashboards, just ask — the Agent understands your workspace data and answers in plain language. ## What you can ask The Agent has context on your visibility, prompts, competitors, sources, and tasks. Useful questions include: * "How did my visibility change this week?" * "Which competitors gained the most share of voice?" * "What's driving my negative sentiment right now?" * "Summarize my top opportunities and what I should do first." ## How it works Open **Agent** from the sidebar and type your question. Answers stream back in real time. Keep the conversation going — the Agent remembers the thread, so you can drill in without repeating yourself. Open a new chat for a new topic. Past chats stay in the sidebar so you can return to them. The Agent has a daily message allowance per workspace. Your remaining messages for the day are shown next to the chat list. Use the Agent to prep for a meeting: ask it to summarize the week's movements and your top recommended actions, then dig into the details on the relevant dashboards. # AI Perception Source: https://docs.kime.ai/features/ai-perception Understand how AI models talk about your brand and what's driving their tone. AI Perception goes beyond a single sentiment score to show you **why** AI models describe your brand the way they do. It's your reputation dashboard for the AI era — track how perception changes over time and pinpoint what's helping or hurting you. ## What you can do Watch your overall sentiment score move and catch shifts early. Identify the factors and themes pushing your perception up or down. See which words and topics drive positive and negative mentions of your brand. Discover which websites AI pulls from when it talks about you, so you know where to focus. ## Filter to compare Break perception down by **country**, **AI model**, **category**, and **time period** to see how your brand is viewed across markets and platforms. Perception often varies significantly from one region or model to another — these filters help you spot where. ## AI summary AI Perception generates a plain-language summary of how models perceive your brand overall, highlighting your key strengths and the areas creating the most friction. It gives you a clear starting point for action without combing through every response. When you spot a source driving negative perception, send it to [Actions](/features/actions) so your team has a concrete task to address it. # Competitor Analysis Source: https://docs.kime.ai/features/competitor-analysis Compare your brand against competitors across AI visibility metrics in depth. Competitor Analysis is your dedicated space for benchmarking. It brings your competitive metrics together so you can see exactly where you lead, where you trail, and where the biggest opportunities are. ## What it shows Compare your brand against your tracked competitors across the core metrics — [visibility](/metrics/visibility), [placement](/metrics/placement), [sentiment](/metrics/sentiment), and [share of voice](/metrics/share-of-voice) — with trends that show who's gaining and who's slipping over your selected period. ## How to use it * **Find the gaps.** Spot categories where a competitor is mentioned and you're not — your clearest opportunities. * **Understand the leaders.** See which brands AI models favor as top recommendations, and in which parts of your category. * **Track momentum.** Watch whether a competitor's recent efforts are moving their numbers, and respond before they pull ahead. Filter by model, market, category, and tags to make every comparison apples-to-apples. Combine Competitor Analysis with [Sources](/features/sources): once you know who's winning a category, check which sources are putting them there. # Prompt Explorer Source: https://docs.kime.ai/features/prompt-explorer Discover the prompts that matter for your brand and start tracking your visibility in AI answers, organized by topic from day one. ### Overview Prompt Explorer helps you find and organize the prompts worth tracking for your brand, without manual research. Based on what your company does, KIME suggests relevant prompts and groups them into topic categories, so you can quickly build a structured tracking setup in your workspace. Each prompt represents a real question a user might ask an AI assistant like ChatGPT, Perplexity, Gemini, or Google AI Overviews. By tracking a prompt, you can see how your brand ranks inside AI answers for that query, and how that changes over time. ### How it works When you create your organization in KIME, you add your company URL as part of setup. Prompt Explorer uses that URL to understand what your company does, including your products, services, and focus areas, and turns that into prompt suggestions. Because your URL is already in place from onboarding, Prompt Explorer works out of the box. There is no extra setup required to start discovering prompts. 1. KIME analyzes your company based on the URL you added when creating your organization. 2. It generates a set of prompts that reflect what your company does and what users in your space are likely to ask AI assistants. 3. Each prompt is sorted into a topic category associated with your business. 4. You browse the suggestions by category and select the prompts you want to track. ### What it's used for Prompt Explorer is built to take the guesswork out of deciding what to track. Use it to: * **Discover relevant prompts** automatically, based on what your company actually does. * **Set up tracking by category**, so your workspace stays structured instead of becoming a flat list of prompts. * **Understand your coverage** across the topics that define your business, and spot where you appear in AI answers and where you don't. * **Scale over time** by adding more prompts as your business grows or shifts focus, while keeping everything organized by topic. ### Setting up prompts by category Prompts come pre-grouped into topic categories tied to your company, which makes building your tracking setup fast and organized: 1. Open your workspace and go to **Prompt Explorer**. 2. Browse the suggested prompts, grouped by topic category. 3. Review each category to see which themes are most relevant to your brand. 4. Select the prompts you want to track. KIME then begins monitoring how your brand ranks inside AI answers for those prompts. As you add more prompts, they stay organized by category, so your team can quickly see brand performance across each area of your business. ### Tips * Start with the categories most central to your business to get meaningful visibility data quickly. * You can also add your own prompts and add tags to each prompt, to make it easier to track your growth across tags and categories. # Sources Source: https://docs.kime.ai/features/sources See which websites and pages AI models cite when they answer about your category. AI assistants build their answers from sources they trust. The **Sources** page shows you which websites and pages those models cite across your tracked prompts — the single most actionable input for improving your AI visibility. ## Why sources matter If you want to change what AI says about your brand, you start with the content it reads. Sources reveals the domains shaping answers in your category, so you can: * Find the high-authority sites that influence recommendations. * See where competitors are mentioned and you're not. * Prioritize the publications, reviews, and pages worth earning a presence on. ## Explore your sources * **Source domains** — browse the websites cited across your prompts, ranked by how often they appear. * **Drill into a domain** — open any source to see the specific prompts and AI responses that referenced it. * **From a prompt** — open a prompt to see the sources behind its answers, then trace each back to the original response. You can filter sources by model, market, category, and tags to focus on a specific slice of your tracking. Build a target list from your most-cited source domains, then use [Actions](/features/actions) to plan how to earn a mention on each one. # Competitors Source: https://docs.kime.ai/getting-started/adding-competitors Choose the rival brands you want to benchmark against in AI answers. AI visibility is competitive: when a model recommends a rival, that's often a slot your brand didn't get. Tracking competitors lets Kime benchmark your performance and calculate metrics like [Share of Voice](/metrics/share-of-voice). Manage your competitive set from the **Competitors** page in the sidebar. Every brand you track here is monitored across all of your prompts. ## Add competitors from suggestions Kime analyzes your domain and proposes brands competing for the same AI visibility. Select **Competitors** in the sidebar to see your current set. Scan the brands Kime identified as your digital competitors. Approve a brand to start tracking it, or reject it to remove it from consideration. ## Add competitors manually Have a specific rival in mind? Click **Add competitor** and enter the brand name and its root domain. The root domain helps Kime attribute mentions accurately across different AI responses. ## Keep your set current Archive competitors from the **Active** tab when they're no longer relevant. Archiving keeps your dashboard focused on the brands you actually compete with and prevents stale brands from cluttering your benchmarks. Track a focused set of genuine competitors. A tight, relevant competitive set produces sharper benchmarks than a long list of loosely related brands. # Answer engines Source: https://docs.kime.ai/getting-started/answer-engines Choose which AI models Kime tracks in each workspace. An **answer engine** is an AI model that responds to your prompts — ChatGPT, Google AI Mode, Gemini, Claude, Perplexity, Copilot, Grok, and more. Kime lets you control which engines are active in each workspace so your data reflects the platforms your customers actually use. ## Select your models Open **Workspace settings** and find the **Answer Engine Selection** panel. Toggle each model on or off: * **On** — the model runs your prompts and contributes to your metrics. * **Off** — the model is excluded from tracking and from your scores. New workspaces start with a sensible default selection, which you can adjust at any time. ## Choosing which to track * **Match your audience.** Prioritize the assistants your customers rely on most in your markets. * **Compare platforms.** Keep several engines active to see where you're strong and where you're invisible — visibility often varies a lot between models. * **Filter later.** On the dashboard you can isolate any single model to analyze its results on their own. Changing which models are active affects future refreshes. Metrics for a newly enabled model build up from the point you turn it on. Keep your most important models consistently enabled so your trend lines stay comparable over time. # Business vs Agency Source: https://docs.kime.ai/getting-started/business-vs-agency Choose the account type that fits how your team works. When you first set up KIME, you choose an account type. It shapes how your workspaces are organized - you can always change it later, so pick what fits today. For brands growing their **own** AI presence. You focus on a single brand that operates in one market or across multiple markets (you have the opportunity to isolate each market in its own workspace). For teams managing AI presence **across many clients**, with isolated workspaces per client and team-based access. ## Business Choose **Business** if you're tracking and improving your own brand. You get the full toolkit: visibility and sentiment across every AI platform, AI Perception, and agentic Actions to close visibility gaps. All Actions in a workspace is personalized to the brand and or market you are tracking. ## Agency Choose **Agency** if you manage AI presence on behalf of clients. In addition to everything in Business, you get: * **Client workspaces** to manage the clients you have - add as many as you need. All agency plans comes with unlimited workspaces. * **Pitch workspaces** for prospects who are curious about AI search, but haven't started yet. * Metrics, perception, and actions for **every client**, kept fully isolated and custom made for each client. * **Role-based access** so teammates and clients see exactly what they should. Learn more in the [Agency](/agency/setting-up-your-first-client) section. Both account types track the same AI models and use the same metrics. The difference is how workspaces and access are organized. # Connect your brand Source: https://docs.kime.ai/getting-started/connect-your-brand Create your workspace and give Kime the brand details it needs to build an accurate tracking profile. Connecting your brand is the first thing you do in Kime. It creates a **workspace** — an isolated space that holds one brand's prompts, competitors, settings, and results. ## Create your workspace Provide the official name of your company. This name is used across the dashboard to identify your primary brand metrics. Use the official name customers know you by. Kime uses this name to identify your brand wherever it appears in AI responses. Input your primary website address, such as `gymshark.com`. KIME requires a root domain rather than specific subfolders or product pages. Enter your primary website, such as `gymshark.com`. Use the root domain rather than a subfolder or product page — it acts as the anchor Kime uses to understand your brand and generate relevant suggestions. This root domain acts as the anchor for KIME, enabling us to scan your website to build an accurate profile of your brand. Kime analyzes your brand and prepares a starting set of prompts and competitors so you don't begin from a blank page. While your workspace is being built, some pages stay locked. Once the build finishes, your full navigation and first results become available. ## Starter suggestions After you connect your domain, KIME automatically proposes: * **Prompts**: questions customers ask AI about in your category and industry. * **Competitors**: brands that operate in the same category as yours and compete for the same AI visibility - you get to benchmark against competitors to see who is most visible in AI answers. Review these in the **Proposed** tabs on the [Prompts](/getting-started/setting-up-prompts) and [Competitors](/getting-started/adding-competitors) pages, then approve the ones that fit and reject the rest - you can also add your own. ## Workspace details You can update your brand information at any time in **Workspace settings**: | Setting | What it controls | | --------------- | --------------------------------------------------- | | Brand name | How KIME identifies your brand in results | | Root domain | The website used to anchor tracking and suggestions | | Market location | The country prompts are run from by default | | Prompt language | The default language for new prompts | Changing your root domain affects how KIME attributes mentions in future refreshes, so update it only when your primary website actually changes. ## Manage who has access Invite teammates from **Workspace settings** and assign roles based on what they need to do. For a full breakdown of each role and what it can do, see [Roles & permissions](/administration/roles-and-permissions). # Introduction Source: https://docs.kime.ai/getting-started/introduction Track and improve how AI models talk about your brand across ChatGPT, Google AI Mode, Gemini, Claude, Perplexity, Copilot, and Grok. Your audience isn't only searching Google anymore. They ask ChatGPT for recommendations, research with Perplexity, compare options in Google AI Mode, and get answers from Gemini, Claude, Copilot, and Grok. People are having conversations with AI instead of typing keywords — and that means brand visibility works differently now. **Kime shows you how your brand appears in AI answers, and what you can do to influence it.** Go from sign-up to your first results in a few minutes. Set up your workspace and let Kime build a tracking profile. Learn what visibility, placement, sentiment, and share of voice mean. Bring your Kime data into Claude, ChatGPT, and other AI tools. ## What Kime does Kime tracks how your brand shows up when people ask AI about your industry. Instead of traditional search rankings, we focus on the things that matter in AI answers: * **Visibility** — how often your brand is mentioned across AI responses. * **Placement** — how prominently you appear when you are mentioned. * **Sentiment** — whether AI describes your brand positively, neutrally, or negatively. * **Share of Voice** — how much of the category conversation you own versus competitors. Underneath all of this are [Sources](/features/sources) — the websites and content AI models reference when they mention brands. When your visibility or placement shifts, it's usually because sources changed what they say about your category, which makes them your biggest opportunity to improve. ## What you'll find in these docs Understand what Kime tracks, pick your setup, and get your first insights fast. Connect your brand and build your tracking: prompts, competitors, markets, and answer engines. Read your headline scores and competitive rankings at a glance. Definitions for every metric and how to interpret them together. Go deeper with AI Perception, Sources, and Competitor Analysis. Turn data into prioritized work with Actions, and ask questions with the Agent. Manage roles, team access, and connect AI assistants to your data. Manage many client brands, run pitch audits, and scale your operation. ## Who uses Kime Kime is built for teams that take AI search seriously — **marketing and brand leaders** tracking positioning, **SEO and GEO teams** treating AI as its own channel, **ecommerce teams** protecting product-category recommendations, and **agencies** managing visibility across many clients. AI answers are non-deterministic and vary day to day, so Kime focuses on trends and patterns over time rather than any single response. See [Data & accuracy](/metrics/other/metric-limitations) for how to interpret your data. # Markets & languages Source: https://docs.kime.ai/getting-started/localizing-prompts Track AI visibility in specific countries and languages to understand how your brand performs in each market. AI assistants tailor their answers to where a user is and the language they speak. They surface local competitors, regional sources, and market-specific recommendations. Tracking from a single market gives you only part of the picture — KIME lets you measure visibility market by market. ## Why location matters The same prompt can produce very different answers in different countries. A brand that dominates AI recommendations in one market may be invisible in another, where regional competitors take its place. Tracking by market shows you exactly what local customers see, so you can: * Confirm you're being recommended in the markets that matter to you. * Spot regional competitors displacing your brand. * Adapt your content strategy to improve where you're underperforming. ## Set a market for a prompt You can choose a market when creating a prompt or when editing an existing one. Find the location setting in the **Add prompt** or **Edit prompt** dialog. Pick the country where your audience is. Kime tracks this prompt as a user in that market would experience it. The prompt begins tracking for that market on the next refresh. Kime currently supports tracking at the country level, so you can compare performance across different regional and cultural markets. ## Write prompts in local language For non-English markets, write prompts the way local customers actually phrase them rather than translating word-for-word. Natural, local phrasing produces far more accurate results. * **Use local terminology.** A UK shopper might ask about "trainers" while a US shopper asks about "sneakers". * **Match real intent.** Reflect how people in that market actually talk to an AI assistant. * **Validate the output.** Use the latest model answers on your dashboard to confirm the localized response reads naturally for that market. You can also set a default **prompt language** for the whole workspace in [Workspace settings](/getting-started/connect-your-brand). Create market-specific versions of your most important prompts so you can compare visibility side by side across the countries you sell in. # Quick Setup Guide Source: https://docs.kime.ai/getting-started/quick-setup-guide Go from sign-up to your first AI visibility results in a few minutes. This guide takes you from sign-up to your first results. The goal: get meaningful, actionable insight into your AI visibility within your first week. Setup itself takes only a few minutes - the rest is KIME collecting data across AI models and helping you and your team on how to optimize your visibility. On first sign-in, pick **Business** (track your own brand) or **Agency** (manage client brands). Not sure? See [Business vs Agency](/getting-started/business-vs-agency). Enter your brand name and root domain. Kime analyzes your brand and builds a starter profile with suggested prompts and competitors. See [Connect your brand](/getting-started/connect-your-brand). Approve the proposed [prompts](/getting-started/setting-up-prompts) and [competitors](/getting-started/adding-competitors) that fit, and add your own. Good prompts mirror the real questions customers ask AI, like *"What's the best CRM for a small marketing team?"* - not keywords. These define what KIME tracks. Confirm which AI models are active for the workspace in [Answer Engine Selection](/getting-started/answer-engines). Choose which models you would like to track and optimize your brands' visibility on. KIME runs your prompts across the selected models and populates your dashboard. Your first full set of scores appears once the initial run completes. Open the [Dashboard](/dashboard/dashboard-overview) to see your visibility, placement, sentiment, and how you rank against competitors. ## What you'll have after setup * A workspace tracking the prompts and competitors that matter to your brand and industry. * Your first visibility, placement, and sentiment scores across the AI models you are tracking on. * A baseline you can build on as you add more prompts and watch trends over time. ## What to do next Use Prompt Explorer to find and add prompts that might be relevant for your brand to track. Dig into how AI models perceive your brand with AI Perception. See which sources AI cites so you know where to focus and which sources matters for your brand to be mentioned on. Work through prioritized recommendations in the Actions board to improve your visibility. # Prompts Source: https://docs.kime.ai/getting-started/setting-up-prompts Build the library of questions KIME tracks across AI models, and organize them with categories, tags, and volume. Prompts are the heart of KIME. Each prompt is a natural-language question — the kind a real customer would type into an AI assistant — that KIME runs across your selected models so it can measure how your brand shows up. ## Prompts vs. keywords Traditional SEO targets keywords that return a list of links. AI assistants instead synthesize a conversational answer, so KIME tracks **prompts**: full questions that capture intent and context, like *"What are the best gym tops for hot-weather training?"* A good prompt mirrors how people *actually* talk to and ask AI. Every effective prompt has three ingredients: * **Subject** — the product, brand, or category in question. * **Context** — details that narrow the answer, like "for marathon training" or "under \$50". * **Intent** — what the user wants: a recommendation, a comparison, or a specific fact. ## Add prompts KIME proposes prompts based on your brand and category. Open the **Proposed** tab on the Prompts page, then **Approve** the ones that fit (they start tracking on the next refresh) or **Reject** the rest. Click **Add prompt**, type the question you want to track, choose the market and category, and save. It joins your active list and is included in the next refresh. Import many prompts at once with a CSV upload. Map your columns to Kime fields during import — you can include a **Tags** column (comma- or semicolon-separated) to tag prompts as they're created. ## Organize your library As your library grows, these tools keep it manageable and your reporting sharp. Group prompts into categories that reflect your business — product lines, comparison queries, problem-solution questions, branded vs. non-branded, and so on. Most dashboards and analysis pages can be filtered by category. Manage categories in **Workspace settings**. Create colored tags to label and filter prompts across the product. Assign them individually, when editing a prompt, in bulk from the prompts table, or via CSV import. Tags appear on prompt rows and can filter the Dashboard, Sources, Competitor Analysis, AI Perception, and more. Manage tags in **Workspace settings**. Each prompt shows a color-coded **Volume** signal that estimates the search demand behind its intent. Use it to prioritize high-volume topics so your optimization effort follows real customer demand, not guesswork. ## Manage your active prompts Your prompt library refreshes automatically on a regular cycle, so your scores keep reflecting the latest AI responses. Use the **Active**, **Proposed**, and **Inactive** views to control what's currently being tracked, and archive prompts you no longer need. Not sure which prompts to add? Open the [Prompt Explorer](/features/prompt-explorer) to browse suggested prompts by topic or category and add them in bulk. # MCP Server Source: https://docs.kime.ai/mcp Connect Kime to Claude, ChatGPT, Cursor, and other AI assistants so you can take action on your AI visibility — work your action plan and launch prompts, right from a chat. Take action on your AI visibility without leaving your AI assistant. The Kime MCP server connects tools like Claude, ChatGPT, Cursor, and Claude Code — or any client that supports the Model Context Protocol (MCP) — directly to Kime, so you can work your action plan, create and update actions, and launch new prompts to track, all from a conversation. You're connected to the same data the dashboard shows, so the context is always there when you need it. Dig into visibility, competitors, sentiment, and sources to understand the story behind an action — then act on it right away. ## What you can do Pull the actions Kime ranks by predicted impact, create new ones, and move them forward — your assistant confirms before it writes. Add prompts in bulk to start tracking, then review how each one performs. See how you show up across every AI model and where you stand against competitors on visibility, placement, and share of voice. Explore sentiment, the keywords driving tone, and the sources AI models cite when they mention your brand. ## Setup guides Launch Claude on your desktop. Go to **Customize** → **Connectors**. Click the **+** sign to add a connector, then select **Add custom connector**. Call it **Kime** and enter this URL: `https://mcp.kime.ai/mcp` Press **Add**, and then **Connect**. The first time you ask Claude about Kime data, it will prompt you to securely sign in. Custom MCP connectors in ChatGPT require **Developer mode** (available on Plus, Pro, Business, Enterprise, and Education plans). On Business and Enterprise workspaces, an admin may need to enable Developer mode first. Open ChatGPT in your browser. Go to **Settings** → **Apps** → **Advanced settings**. Turn on **Developer mode** and accept the warning. Click **Create app** (this only appears when Developer mode is on). Name it **Kime**, paste this URL, and choose **OAuth** as the authentication method: `https://mcp.kime.ai/mcp` Save the app, then use **Scan tools** or **Refresh** to confirm ChatGPT can reach Kime. In a new chat, open the **+** menu → **Developer mode**, and enable your Kime connector. The first time ChatGPT uses a Kime tool, it will prompt you to securely sign in. Kime works with any MCP-compatible client over Streamable HTTP. Point it at the connection URL and sign in via OAuth when prompted. ```bash Claude Code theme={null} claude mcp add kime --transport http https://mcp.kime.ai/mcp ``` ```json Cursor (mcp.json) theme={null} { "mcpServers": { "kime": { "url": "https://mcp.kime.ai/mcp" } } } ``` ```json VS Code (.vscode/mcp.json) theme={null} { "servers": { "kime": { "type": "http", "url": "https://mcp.kime.ai/mcp" } } } ``` In **Cursor**, you can also open **Settings → Tools & Integrations → MCP**, add a custom MCP server with the Kime URL, and select Streamable HTTP. For full setup steps and troubleshooting across every client, see [Connect AI assistants](/administration/integrations). ## Example questions Ask in plain language. Here are some prompts to get you started: **Visibility & performance** * "How has our visibility changed over the last 30 days?" * "Which AI models mention us the most?" * "Compare our visibility in ChatGPT versus Gemini this month." **Competitors** * "Which competitors have the highest share of voice?" * "Where do competitors show up but we don't?" * "Compare our placement against \[competitor] over the last 14 days." **AI perception** * "What's our sentiment score across AI models right now?" * "Break our sentiment down by factor and show the strongest negative excerpts." * "Which keywords drive negative sentiment about our brand?" * "Summarize how AI models currently describe us." **Sources & citations** * "Which domains get cited most when our brand comes up?" * "What sources are behind the most positive mentions of us?" **Take action** * "What should we work on next to improve our AI visibility?" * "List our open actions and create one to publish a comparison page versus \[competitor]." * "Mark the \[competitor] comparison action as in progress and add a note with the angle." * "Add these 10 prompts and start tracking them." ## Authentication Kime uses **OAuth** — there's nothing to paste or store. The first time your assistant uses a Kime tool, it opens a secure sign-in window where you log in and approve access. After that, your session persists across conversations. Access always respects your Kime account: an assistant can only see the organizations and workspaces you can, and acts within your existing permissions. The consent screen identifies which application is requesting access, and you can switch accounts if needed. ## Read and write access Most tools are **read-only**. Visibility, sentiment, sources, competitor benchmarks, prompts, and actions all pull existing data without changing anything. A smaller set of tools help you act on what you find: * **Work your actions** — pull the prioritized action plan Kime surfaces, create new actions, update their status, and add comments as you make progress. * **Create prompts** — add prompts in bulk and start tracking them across your models. Write tools are clearly flagged, so your client asks for confirmation before anything runs. You grant the matching permissions when you connect, and changes always stay within your existing Kime access. ## Supported platforms The Kime MCP server works with any client that supports the Model Context Protocol, including Claude, ChatGPT, Cursor, Claude Code, VS Code (GitHub Copilot), and Windsurf. ## Requirements A Kime account with at least one active workspace, and an MCP-compatible AI tool. # MCP Server Source: https://docs.kime.ai/mcp-server Connect Kime to Claude, ChatGPT, Cursor, and other AI assistants so you can take action on your AI visibility — work your action plan and launch prompts, right from a chat. Take action on your AI visibility without leaving your AI assistant. The Kime MCP server connects tools like Claude, ChatGPT, Cursor, and Claude Code — or any client that supports the Model Context Protocol (MCP) — directly to Kime, so you can work your action plan, create and update actions, and launch new prompts to track, all from a conversation. You're connected to the same data the dashboard shows, so the context is always there when you need it. Dig into visibility, competitors, sentiment, and sources to understand the story behind an action — then act on it right away. ## What you can do Pull the actions Kime ranks by predicted impact, create new ones, and move them forward — your assistant confirms before it writes. Add prompts in bulk to start tracking, then review how each one performs. See how you show up across every AI model and where you stand against competitors on visibility, placement, and share of voice. Explore sentiment, the keywords driving tone, and the sources AI models cite when they mention your brand. ## Setup guides Launch Claude on your desktop. Go to **Customize** → **Connectors**. Click the **+** sign to add a connector, then select **Add custom connector**. Call it **Kime** and enter this URL: `https://mcp.kime.ai/mcp` Press **Add**, and then **Connect**. The first time you ask Claude about Kime data, it will prompt you to securely sign in. Custom MCP connectors in ChatGPT require **Developer mode** (available on Plus, Pro, Business, Enterprise, and Education plans). On Business and Enterprise workspaces, an admin may need to enable Developer mode first. Open ChatGPT in your browser. Go to **Settings** → **Apps** → **Advanced settings**. Turn on **Developer mode** and accept the warning. Click **Create app** (this only appears when Developer mode is on). Name it **Kime**, paste this URL, and choose **OAuth** as the authentication method: `https://mcp.kime.ai/mcp` Save the app, then use **Scan tools** or **Refresh** to confirm ChatGPT can reach Kime. In a new chat, open the **+** menu → **Developer mode**, and enable your Kime connector. The first time ChatGPT uses a Kime tool, it will prompt you to securely sign in. Kime works with any MCP-compatible client over Streamable HTTP. Point it at the connection URL and sign in via OAuth when prompted. ```bash Claude Code theme={null} claude mcp add kime --transport http https://mcp.kime.ai/mcp ``` ```json Cursor (mcp.json) theme={null} { "mcpServers": { "kime": { "url": "https://mcp.kime.ai/mcp" } } } ``` ```json VS Code (.vscode/mcp.json) theme={null} { "servers": { "kime": { "type": "http", "url": "https://mcp.kime.ai/mcp" } } } ``` In **Cursor**, you can also open **Settings → Tools & Integrations → MCP**, add a custom MCP server with the Kime URL, and select Streamable HTTP. For full setup steps and troubleshooting across every client, see [Connect AI assistants](/administration/integrations). ## Example questions Ask in plain language. Here are some prompts to get you started: **Visibility & performance** * "How has our visibility changed over the last 30 days?" * "Which AI models mention us the most?" * "Compare our visibility in ChatGPT versus Gemini this month." **Competitors** * "Which competitors have the highest share of voice?" * "Where do competitors show up but we don't?" * "Compare our placement against \[competitor] over the last 14 days." **AI perception** * "What's our sentiment score across AI models right now?" * "Break our sentiment down by factor and show the strongest negative excerpts." * "Which keywords drive negative sentiment about our brand?" * "Summarize how AI models currently describe us." **Sources & citations** * "Which domains get cited most when our brand comes up?" * "What sources are behind the most positive mentions of us?" **Take action** * "What should we work on next to improve our AI visibility?" * "List our open actions and create one to publish a comparison page versus \[competitor]." * "Mark the \[competitor] comparison action as in progress and add a note with the angle." * "Add these 10 prompts and start tracking them." ## Authentication Kime uses **OAuth** — there's nothing to paste or store. The first time your assistant uses a Kime tool, it opens a secure sign-in window where you log in and approve access. After that, your session persists across conversations. Access always respects your Kime account: an assistant can only see the organizations and workspaces you can, and acts within your existing permissions. The consent screen identifies which application is requesting access, and you can switch accounts if needed. ## Read and write access Most tools are **read-only**. Visibility, sentiment, sources, competitor benchmarks, prompts, and actions all pull existing data without changing anything. A smaller set of tools help you act on what you find: * **Work your actions** — pull the prioritized action plan Kime surfaces, create new actions, update their status, and add comments as you make progress. * **Create prompts** — add prompts in bulk and start tracking them across your models. Write tools are clearly flagged, so your client asks for confirmation before anything runs. You grant the matching permissions when you connect, and changes always stay within your existing Kime access. ## Supported platforms The Kime MCP server works with any client that supports the Model Context Protocol, including Claude, ChatGPT, Cursor, Claude Code, VS Code (GitHub Copilot), and Windsurf. ## Requirements A Kime account with at least one active workspace, and an MCP-compatible AI tool. # Benchmarking Source: https://docs.kime.ai/metrics/other/benchmarking How to evaluate your AI visibility metrics against industry standards and competitors. AI visibility is still a young field, and what counts as a "good" score varies widely by industry. Benchmarking gives your numbers meaning by comparing them to a baseline and to your competitors. ## Establish a baseline Spend your first few weeks watching your scores settle before setting targets. This baseline shows your brand's natural range and gives you a realistic starting point for measuring growth. ## Benchmark against competitors The most useful comparison is against your direct rivals. Use the [Industry ranking](/dashboard/industry-ranking) table to line up your metrics with your competitive set: * **Visibility** — if the leader is far ahead, find the categories where they out-mention you and focus there. * **Placement** — if rivals are consistently positioned ahead of you, your content may not yet read as authoritative enough to be the top pick. * **Sentiment** — compare your tone to the category to see whether reputation is an advantage or a liability. ## Benchmark by category Always benchmark within your prompt categories. A brand might dominate branded queries while barely appearing in broader category questions. Looking category by category tells you whether you're only visible to people who already know you — or genuinely winning new audiences through AI. Set targets relative to your competitors and your own trend, not to an absolute number. Beating last month and closing the gap on the leader are the goals that matter. # Reading your metrics Source: https://docs.kime.ai/metrics/other/how-we-calculate-metrics How to interpret Kime's scores together to understand your brand's health in AI answers. Kime turns thousands of real AI responses into a small set of clear scores. This page explains how to **read** those scores together — each one answers a different question, and they're most powerful in combination. ## What each score answers | Metric | The question it answers | | ----------------------------------------- | ----------------------------------------------- | | [Visibility](/metrics/visibility) | How often does AI mention my brand? | | [Placement](/metrics/placement) | How prominently am I positioned in answers? | | [Sentiment](/metrics/sentiment) | How positively is my brand described? | | [Share of Voice](/metrics/share-of-voice) | How much of the category conversation do I own? | ## The AI Performance Score The **AI Performance Score** is a single composite index of overall brand health. It blends how often you're mentioned, how prominently you're positioned, and how positively you're described into one number you can track over time. Because it combines all three dimensions, no single metric can paint a misleading picture — strong visibility won't inflate your score if sentiment is poor. ## Read them as a set The real insight comes from how the scores move together: * **Visibility up, placement flat** — you're getting mentioned more, but not yet recommended first. * **Visibility up, sentiment down** — more attention, but not the right kind; investigate the tone. * **SoV down, visibility steady** — competitors are growing faster than you. * **Performance score rising** — your overall position is strengthening across the board. Always pair a headline number with a breakdown. Filtering by model, market, and category turns a single score into a specific, actionable insight. # Data & accuracy Source: https://docs.kime.ai/metrics/other/metric-limitations What Kime's metrics can and can't tell you, so you can interpret them with confidence. Kime gives you a consistent, comparable way to measure AI visibility. As with any measurement of AI behavior, it helps to understand what the data represents — and what it doesn't. ## AI answers vary AI models are non-deterministic: the same question can produce slightly different answers each time. Kime's metrics are designed as a reliable strategic signal — a high-level view of trends and patterns — rather than an exact count of every mention across the entire AI ecosystem. ## Results differ by model Your scores reflect the specific prompts and models you track. Because each model is trained and updated differently, your performance will naturally vary between them. * **Model updates** — a sudden shift can come from a model changing how it answers, not from anything you did. * **Live web results** — models that browse the web in real time may surface different sources from one moment to the next. * **Market differences** — results are tracked at the country level, so very local variations may differ from what Kime reports. ## Patterns over precision AI models don't publish ranking factors the way search engines describe theirs. Kime can show you **where** you appear, **how** you're positioned, and the **tone** of each mention — use these to spot patterns and correlations and to guide strategy, rather than treating any single number as an exact, fixed measurement. ## A strategic tool Metrics refresh on a regular cycle, so the dashboard reflects daily, weekly, and monthly trends rather than live, second-by-second activity. That makes Kime ideal for steering long-term content and optimization strategy and measuring its impact over time. When a number moves sharply, check whether it lines up across several brands. A category-wide shift usually points to a model change rather than something specific to your brand. # Placement Source: https://docs.kime.ai/metrics/placement How prominently your brand is positioned within AI answers. **Placement** measures how prominently your brand appears within an AI answer. AI responses are conversational rather than a ranked list of links, so placement reflects how early and how prominently your brand shows up relative to others. A stronger placement means models are treating your brand as a leading recommendation. ## Why it matters Being mentioned is only half the battle — being mentioned **first** is what earns attention and trust. Most people focus on the top one or two options an AI suggests. Strong visibility paired with weak placement usually means models know your brand but see competitors as the more authoritative answer. ## How to read it Use placement to understand your authority in a category: * **Leading** — your brand is the primary recommendation for most queries. * **Mid-tier** — you're frequently mentioned, but often after a more dominant competitor. * **Trailing** — you appear as an alternative or honorable mention rather than a primary pick. ## How to improve it Improving placement is about authority, not just presence. Aim to be featured in the kind of high-trust, "best of" and comparison content that models lean on when deciding which brands to put first — and make sure that coverage presents your brand as a top choice. If your visibility is strong but placement is weak, prioritize earning standout coverage in authoritative sources over simply getting more mentions. # Sentiment Source: https://docs.kime.ai/metrics/sentiment The tone AI models use when they describe your brand. **Sentiment** measures the tone of an AI answer when it mentions your brand. Where visibility tracks how often you appear, sentiment tracks **how you're perceived**. Each mention is understood as positive, neutral, or negative, giving you a clear read on your reputation inside AI answers. ## The categories * **Positive** — the AI recommends your brand, highlights benefits, or describes it favorably. * **Neutral** — your brand is mentioned factually, such as in a list, without a clear opinion. * **Negative** — your brand comes up in the context of complaints, limitations, or unfavorable comparisons. ## The sentiment score Kime rolls these mentions into an overall **sentiment score**, where higher means a more positive presence. Tracking it over time shows whether your reputation in AI answers is improving or slipping, and you can break it down by model and market to see where perception differs. ## How to use it Sentiment gives your other metrics context. High visibility is only valuable if the tone is positive — lots of mentions paired with a low sentiment score means AI models are highlighting problems, often pulling from dated reviews or negative coverage. For a deeper view of what's driving your tone — the factors, keywords, and sources behind it — use [AI Perception](/features/ai-perception). Watch visibility and sentiment together. Rising visibility with falling sentiment is a signal to address reputation issues before they spread further across AI answers. # Share of Voice Source: https://docs.kime.ai/metrics/share-of-voice The slice of the AI conversation your brand owns versus your competitors. **Share of Voice (SoV)** is a competitive metric: it tells you how much of the conversation in your category your brand owns compared to your competitors. Where [visibility](/metrics/visibility) measures your presence across your own prompts, SoV measures your presence relative to everyone you track. A high SoV means that, among the brands competing in your category, yours captures a large portion of the mentions AI models make. ## Why it matters for AI Space in an AI answer is limited — models typically name only a handful of brands per response. A strong SoV means you're consistently winning those scarce slots. If your SoV is falling while your visibility holds steady, it's a sign competitors are growing faster and starting to crowd you out. ## Analyze it by category Filter by category to see SoV at a granular level. You might own the conversation in one product line while a competitor dominates another. That breakdown shows you exactly where you have an advantage to defend and where you need to invest. Use SoV alongside the [Mentions leaderboard](/dashboard/mentions-leaderboard) to turn "who gets mentioned" into "who owns the category." # Visibility Source: https://docs.kime.ai/metrics/visibility How present your brand is across the AI answers you track. **Visibility** measures how present your brand is in AI answers. It tells you how much of your tracked prompt library surfaces your brand, expressed as a percentage so you can compare it cleanly across models, markets, and time. A higher visibility score means AI models recognize your brand and bring it up across more of the questions your customers ask. ## How to read it Visibility is most useful in context, not in isolation: * **Versus competitors** — compare your visibility to your competitive set to gauge how dominant you are in the category. * **Across models** — you may be highly visible in one model and nearly absent in another, which points to where you have room to grow. * **By category** — filter by category to see whether your visibility is concentrated in a few product lines while others lag. ## How to improve it AI models build their answers from the content and sources they trust. To raise your visibility, focus on becoming a more common and credible part of that landscape: * Earn mentions on high-authority sites, industry publications, and respected review platforms. * Make sure your brand is clearly represented in the content people reference when researching your category. * Track your progress over time — visibility responds gradually as the broader content ecosystem updates. Pair visibility with [Placement](/metrics/placement) and [Sentiment](/metrics/sentiment). Being mentioned is the first step; being mentioned prominently and positively is what wins customers. # Actions Source: https://docs.kime.ai/untitled-page Manage your LLM optimization roadmap through our AI-driven task pipeline. The **Actions** board is where Kime transforms AI search data into a structured execution plan. Think of it as your **AI CMO**—prioritizing the specific optimizations needed to boost your brand’s visibility and citation confidence across Large Language Models. While Kime delivers end-to-end insights into your performance, the Actions board bridges the gap between raw data and real-world execution. ## The Task Pipeline Moving a task through the pipeline ensures every piece of content or editorial change is validated against LLM best practices. This column contains all pending actions generated by Kime's AI analysis. Each task is prioritized based on its potential impact on your visibility. Start here to select your next high-impact move. Once a team member is assigned, move the card here. This signals to the rest of the organization that the task—whether it's building a brand visibility hub or drafting a Wikipedia article—is currently being executed. After completing the "Must Haves" listed in the task description, move the card here. This stage is critical for checking that content aligns with semantic matching requirements and LLM citation signals before going live. Tasks moved here are considered complete. Kime archives these actions so you can maintain a historical record of your brand’s optimization journey and track velocity over time. ## Inside an Action Every task card contains a deep dive into the "Why" and "How" of LLM optimization. Click any card to reveal the execution blueprint. ### Strategic framework Kime doesn't just tell you *what* to do; it provides the rationale: * **Why it works**: A breakdown of how the action influences LLM retrieval, URL tokens, and semantic match confidence. * **Must haves**: A mandatory checklist of requirements to ensure the task meets KIME's optimization standards. * **Could haves**: Optional enhancements, like visual sitemaps or competitor comparisons, to further boost performance. * **Competitor examples**: Direct links to successful executions by competitors to serve as benchmarks. ### Assignment & Metadata Use the metadata headers to quickly reassign tasks or adjust priority levels as your strategy evolves. | Field | Description | | :----------- | :------------------------------------------------------ | | **Assignee** | The team member responsible for the task. | | **Category** | Classification (e.g., Editorial, Content, Technical). | | **Priority** | The urgency level (Urgent, High, Medium, Low). | | **Created** | The date Kime identified this optimization opportunity. | ## Refining Your View As your roadmap grows, use the top navigation bar to filter the board and focus on what matters most: * **Filter by Assignee**: See only your tasks or check the workload of a specific team member. * **Priority & Favorites**: Quickly access critical "Urgent" tasks or starred items. * **Reject & Delete**: If a task no longer aligns with your immediate brand goals, use the **Action** menu within the card to remove it from the board. **Collaborate in Real-Time:** Use the **Comments** section at the bottom of any task card to share drafts or ask for feedback. Press `Cmd + Enter` to quickly post your comment.