Terminology

AmICited uses a small vocabulary consistently across the dashboard, the MCP server and these docs. Terms are grouped by topic. Where a label in the app differs from the plain meaning, the entry says so.

Core objects#

• Workspace. Your FlowHunt workspace: one team, one billing account and one credit balance shared by AmICited and FlowHunt. Members have roles (Owner, Admin, Editor, Member, Guest, Maintainer). See team and roles.
• Domain. The unit of tracking: one website, with its homepage URL and brand name. Prompts, competitors, tags and every metric belong to one domain; there are no separate projects. See add a domain.
• Brand name. How your brand appears in AI answers, auto-filled from the URL and editable. Together with the domain, it decides which mentions count as you. See add a domain.
• Prompt. A question a buyer might ask an AI assistant, tracked for one domain in one country on a schedule. The AI-search equivalent of a keyword. See choosing the right prompts.
• Engine. An AI system AmICited asks your prompts, such as ChatGPT, Perplexity or Google AI Overview. Also called a provider or platform in the app. See AI engines.
• Response. One engine’s answer to one prompt on one run, stored in full with its citations and brand mentions. Also the billing unit: one response costs 0.015 credits. See how AmICited collects data.
• Run. One scheduled execution of a prompt across its engines. A run on five engines produces five responses. See run frequency.
• Competitor. A brand you choose to track against, with a name and an optional domain. Tracked competitors are pinned in your charts; they do not change how shares are calculated. See competitors.
• Detected brand. Any brand an engine named in your answers, found automatically whether or not you track it. The Competitors page lists them so you can add the ones that matter. See competitors.

Metrics#

• Brand mention. A brand named in the text of an answer, linked or not. Each brand counts once per answer, at its first position. See mentions vs. citations.
• Citation. A URL an engine listed as a source for its answer, with its position in the source list. See mentions vs. citations.
• Visibility score. Prompts where your brand was named ÷ tracked prompts × 100. The headline coverage metric. See visibility score.
• Share of voice. Your brand mentions ÷ all brand mentions × 100, counting every brand named, not only tracked competitors. See share of voice.
• Average rank. The mean position of your brand in the answers that named it, where 1 is named first. Labelled Avg citation rank on the dashboard, although it measures the brand name, not a cited page. See average rank.
• Share of citation. A domain’s or URL’s citations ÷ all citations × 100. Shows whose pages the engines build answers from. See share of citation.
• Not ranked. Tracked prompts where no answer named your brand in the period. Also a prompt status and a filter on the Prompts page. See not-ranked prompts.
• Cited (prompt status). A prompt where at least one answer in the period named your brand. The opposite of not ranked. See prompt detail.
• Total citations. The number of answers that named your brand in the period. Despite the label, it counts mentions. See metrics overview.
• Tracked prompts. The number of prompts matching the current filters; the denominator of visibility. See reading the dashboard.
• Prompt visibility. On a single prompt’s page, your mentions ÷ all brand mentions for that prompt: a share of voice scoped to one question. See prompt detail.
• Change (pts). Every metric is compared with the previous period of the same length. Percentages change in percentage points: 35.0% to 42.5% is +7.5 pts. See reading the dashboard.

Prompt setup#

• Country. Where a prompt is asked from, one of 139 supported countries (default United States). The same text can be tracked in several countries. See countries and languages.
• Language. Derived from the country, not chosen per prompt: each country maps to one primary language. See countries and languages.
• Schedule. How often a prompt re-runs: hourly, every 3 hours, daily (default), weekly, monthly, yearly or one time. More frequent schedules use more credits. See run frequency.
• Engine allow-list. An optional per-prompt list of engines. Left empty, the prompt runs on every enabled engine, including engines added later. See AI engines.
• Premium engines. Claude, Grok and DeepSeek, available on Premium and Enterprise. See AI engines.
• Tag. A label (up to 50 characters, with a color) you give a prompt. One tag per prompt; every dashboard figure can be filtered by tag. See tags.
• Keyword data. Search volume, CPC and competition (0 to 100) for a prompt’s text, used to prioritize prompts. See adding prompts.
• Intent. An automatic label for what the person asking wants: Informational, Navigational, Commercial, Transactional, Local, Exploratory, Comparative research or Synthesis. Shown as a column and filter on the Prompts page. See choosing the right prompts.
• Prompt status. Where a prompt is in its run cycle, shown on the Prompts page as queued, processing or failed. A failed run is retried after 5 minutes. See missing engine data.

Analysis#

• Query fan-out. The sub-queries an engine searches while composing its answer. Reported by ChatGPT and Gemini, predicted for other engines, and grouped into categories. See query fan-out.
• Sources. The page listing every domain and URL cited in your answers, with share, change, engines and a 14-day trend. See Sources.
• Semantic map. A plot of the prompt, its fan-out queries, cited URLs and page headings by meaning, clustered to show topic groups. Built per prompt, per domain and for ChatGPT Ads. See semantic maps.
• Prompt coverage. Suggests which of your pages should own each prompt and marks each fan-out question as covered, partial or uncovered. Three or more uncovered questions in a cluster flag a new-page candidate. See prompt coverage.
• Citation gap. A report answering “Which prompts cite everyone except us?”, joining your organic rank with your citation rate. See citation gap report.
• Ads and products in answers. The share of a prompt’s runs whose answer showed sponsored ads or product listings. See ads and products in AI answers.
• AI searches. A report of Search Console queries containing a site: operator, treated as searches run by AI assistants. See search and traffic reports.
• Freshness sensitivity. A 0 to 100 score for how fresh the top Google results for a prompt are. A high score suggests the prompt rewards recently updated pages. See freshness and Web Vitals.
• Web Vitals sensitivity. A 0 to 100 score for how many of the top Google results for a prompt have good Core Web Vitals. It describes the result set, not a cause of ranking. See freshness and Web Vitals.

Site and content#

• AI article. A draft written to target a prompt, in formats such as blog post, listicle, comparison, FAQ or how-to, exportable as Markdown, HTML, PDF or DOC. See AI articles.
• Annotation. A logged change to your site whose outcome AmICited measures, reported as met, missed or inconclusive. See annotations.
• Domain audit. Checks whether AI agents can reach and read your site: robots.txt, llms.txt review and the accessibility tree. See domain audit.
• llms.txt. A plain-text file at your site root that points AI systems to your key pages. The domain audit reviews yours. See domain audit.
• Site audit. A per-URL crawl of your own site. See domain audit.
• Freshness. How often you and your competitors add and remove URLs, read from sitemaps. See freshness and Web Vitals.
• Web Vitals. Real-user page speed from Chrome UX Report field data. See freshness and Web Vitals.
• Competitor sitemaps. Competitors’ sitemap page counts, 30-day change and freshness. See competitors.
• Uptime monitor. A check that your site is up, with heartbeats, maintenance windows, status pages, an SLA report and alerts by email, Slack or webhook. See uptime monitoring.

Engines and data#

• Google Search baseline. The plain organic results collected alongside AI Overviews. Visible on a prompt’s page, excluded from every citation metric. See how AmICited collects data.
• Data warehouse. Where every response, citation and mention is stored per domain, next to your search, ads and store data. Full history, not sampled. See how AmICited collects data.
• Citation history. How far back your plan lets you view AI answer data, from 7 days on the trial to unlimited on Enterprise. See plans and limits.
• Data source. A connected external platform: Search Console, Bing Webmaster Tools, GA4, an ad platform or an ecommerce store. See Search Console and Bing.
• Credit. The billing unit, worth €1. Responses, keyword lookups, audits and other paid actions draw from the workspace balance. See credits.
• Credit-blocked. The state of a domain when the workspace runs out of credits: scheduled runs pause until credits are added. See credits.

AI agents (MCP)#

• MCP server. The always-on endpoint that gives Claude, ChatGPT, Cursor, Codex and other agents access to your AmICited data. Every dashboard feature is also an MCP tool. See connect the MCP server.
• Toolset. One of 12 product areas the MCP tools are grouped into, such as visibility, prompts or uptime. You can pin toolsets on the server URL. See toolsets and tools.
• Playbook skill. A step-by-step guide served to agents at skill://<name>/SKILL.md, for questions that need several tools in a set order. See playbook skills.
• Agent Mode. A chat inside the app for asking questions about your data, available on Pro and above. See product map.