Every tool Hi, Moose MCP offers a connected app, with what it takes and what comes back. You do not call tools yourself: the app picks the tool from what you ask. This page tells you what to ask for and what to expect.
Every tool except list_projects and get_result takes a project, by the name shown in Hi, Moose. Say the project name in your request when you have more than one, for example "for the Acme project".
How long things take¶
| Tool | Typical time |
|---|---|
| List projects, briefs, drafts | A few seconds |
| Check site coverage | Under a minute |
| Find site pages | About a minute; a few minutes on a large site |
| Visibility check, read an answer | A few seconds |
| Opportunities list | A few seconds to a minute |
| Competitor gap analysis | One to two minutes |
| Tracked list, visibility summary, results list | A few seconds |
| Content brief, draft, FAQs, key points, featured image | One to three minutes |
| AEO audit | Two to four minutes |
| Content refresh | Three to six minutes |
| Internal links | About a minute |
| Audio episode | 5 minutes for short, 9 for standard, 15 for deep |
When a tool takes longer than the app will wait, it answers "still working" with a job id. The app then uses get_result to pick the finished work up; ask it to check again if it does not.
Projects and Library¶
list_projects¶
Lists the Hi, Moose projects you can work on. Every other tool needs a project, so the app usually calls this first.
Ask: "List my Hi, Moose projects."
list_briefs¶
Lists the content briefs saved in the Library for a project, newest first, up to 50. Use it when you refer to a brief made earlier.
Ask: "Show the briefs saved for Acme."
list_drafts¶
Lists the drafts saved in the Library for a project, newest first, up to 50.
Ask: "Which drafts do I have in Acme?"
get_result¶
Picks up work started by another tool, by job id. Waits up to a minute for it to finish. The app calls this on its own; you only need it when the app says the work is still running and you want it to check again.
Site coverage¶
check_site_coverage¶
Checks the latest crawl of your site for pages that already cover one or more topics.
| Input | Required | Meaning |
|---|---|---|
| topics | yes | One to fifteen short phrases, such as "SBOM generation for Python" |
| parentTopic | no | The wider subject the topics belong to, used to judge partial coverage |
For each topic it answers covered, partial, uncovered or unknown, with the closest page URLs as evidence and the date of the crawl it checked. Unknown means the check could not run for that topic, not that the site lacks it.
Ask: "Does our site already cover SCA vs SAST?"
find_site_pages¶
Lists every page in the latest crawl of your site that is related to a topic, for a content inventory or an audit of what already exists. Hi, Moose scores every indexed page against the topic, then a model checks each candidate and drops pages that only share a word with it.
| Input | Required | Meaning |
|---|---|---|
| topic | yes | A short phrase, such as "SBOM generation", up to 200 characters |
| limit | no | A shorter list, when you want only the closest pages. Without it you get every related page, up to 500. |
| urlContains | no | Only pages whose URL contains this text, for example "/learn/" |
| includePassages | no | Also return the passage of each page that best matched the topic. Makes the result much longer. Default false. |
Returns the related pages closest first, each with URL, title, similarity score, page type and dates, plus how many pages were checked, whether the list was cut, and the date of the crawl. Use check_site_coverage instead when the question is whether a topic is covered.
Ask: "List every page we have about SBOM generation."
Content¶
create_content_brief¶
Creates a search-informed content brief: objective, audience, rationale, keyword buckets, recommended structure and constraints, grounded in live search results, your site crawl, brand guidelines and Brand Truth.
| Input | Required | Meaning |
|---|---|---|
| topic | yes | The topic, question or working title, up to 400 characters |
| targetUrl | no | The page on your site the brief is for |
| templateType | no | The kind of content: "Blog post", "Landing page", "Comparison page", "Guide" |
| guidance | no | Requirements decided elsewhere: angle, must-include points, tone. Up to 4,000 characters. |
| opportunityId | no | The id of a row from visibility_list_opportunities. Hi, Moose on the host device looks the row up, adds its gap evidence to the brief and reads the pages the AI engines cited for it. |
| referenceUrls | no | Up to 6 page URLs the AI engines cited for this topic, usually the sources of that opportunity row. Hi, Moose reads each page in full and gives it to the brief as a reference source. Works on the host device and in the cloud. |
Returns the brief as text plus a brief id the draft tool uses. Saved in the Library. When the brief starts from an opportunity, its "Data used" list says how many of those cited pages were read, with word counts.
Ask: "Write a content brief for a guide to SBOM generation for Python teams."
create_draft¶
Writes a publishable first draft from a brief, in your brand's voice, grounded in the brief, Brand Truth and brand guidelines, with internal links chosen from your site crawl and a polish pass.
| Input | Required | Meaning |
|---|---|---|
| briefId | one of the two | A brief made in Hi, Moose |
| brief | one of the two | A brief written elsewhere, as text, up to 60,000 characters |
| briefTitle | no | The working title, when passing a brief as text |
| templateType | no | The kind of content, when passing a brief as text |
| guidance | no | Extra requirements, up to 4,000 characters |
Returns the draft as text plus a draft id. Saved in the Library.
Ask: "Draft the article from that brief." or "Here is a brief I wrote, draft it for Acme: ..."
refresh_content¶
Refreshes content you already have. Hi, Moose reads the current content, works out the search query it targets (or uses the one you give), runs the same live research as a content brief for that query, and compares the two: what still works, what reads as stale, facts to update, what ranking and cited pages cover that your content does not, and the questions searchers ask that it does not answer.
| Input | Required | Meaning |
|---|---|---|
| url | one of the three | The live page. Best, because the comparison then uses exactly what is published. |
| draftId | one of the three | A saved draft |
| content | one of the three | The current content as text, up to 60,000 characters |
| title | no | The title, when passing content as text |
| targetQuery | no | The question or search prompt the refreshed content should win, up to 200 characters. Without it Hi, Moose detects the main query from the content and names it in the result. |
| templateType | no | The kind of content, such as "Blog post" or "Landing page" |
| guidance | no | Requirements for the refresh: sections to keep, points to add, tone. Up to 4,000 characters. |
Returns two things, both saved in the Library: a refresh brief (the comparison, a change list, and a full content brief for the updated piece) and a revised draft that keeps the current structure, voice and links and applies the changes. Download the brief from the Library as PDF, Word or Markdown when you need to make the case for the refresh internally.
Ask: "Refresh https://acme.example/guides/sbom." or "This article is two years old, refresh it for 'SBOM generation tools': ..."
create_featured_image¶
Creates a featured image for an article, 1200 by 630, in your saved image style when one is set.
| Input | Required | Meaning |
|---|---|---|
| draftId | one of the two | A draft made in Hi, Moose |
| title and summary | one of the two | The article title and a short summary, up to 6,000 characters, when there is no draft |
| styleIntent | no | A wish for this one image: mood, scene, colours. The saved style still applies. |
Returns the image, a link to the PNG and alt text written from the finished picture: it describes the image first and names the article's topic once, about 125 characters, in the article's language. The file is also saved on the host device.
Ask: "Make a featured image for that draft, evening light, no people."
create_faqs¶
Writes frequently asked questions with answers for one piece of content, in your brand's voice, plus FAQPage schema markup ready to paste into the page. On paid accounts the questions are researched from real search behaviour first.
| Input | Required | Meaning |
|---|---|---|
| draftId | one of the two | A saved draft |
| url | one of the two | A published page on your site |
| guidance | no | Audience, questions that must appear, tone. Up to 4,000 characters. |
Ask: "Write FAQs for https://acme.example/guides/sbom."
create_key_points¶
Writes the key points of one piece of content: short, quotable statements in your brand's voice that answer engines can lift.
| Input | Required | Meaning |
|---|---|---|
| draftId | one of the two | A saved draft |
| url | one of the two | A published page on your site |
| guidance | no | Audience, points that must appear, tone. Up to 4,000 characters. |
Ask: "Give me key points for the SBOM guide draft."
Audits and links¶
run_aeo_audit¶
Runs an answer engine optimization audit on one piece of content: a score out of 100, category scores (structure, extractability, citation potential, trust signals, freshness, answer readiness and more), the issues found, and concrete changes with proposed text.
| Input | Required | Meaning |
|---|---|---|
| url | one of the three | A live page. Best, because technical checks such as AI crawler access and page dates run only on a live page. |
| draftId | one of the three | A saved draft |
| content | one of the three | The content as text, for a page that is not live yet. Up to 50,000 characters. |
| targetQuery | no | The question or search prompt the content should win, up to 200 characters. Without it Hi, Moose infers one. |
On paid accounts live search evidence is included.
Ask: "Audit https://acme.example/guides/sbom for the query 'SBOM generation tools'."
find_internal_links¶
Plans internal links for one piece of content using the crawl of your own site. Returns two lists: pages this content should link out to, each with the anchor text to use, and existing pages that should add a link pointing to this content, each with the anchor text to use on that page. Pages already linked either way are left out.
| Input | Required | Meaning |
|---|---|---|
| url | one of the three | A page on your site, live |
| draftId | one of the three | A saved draft |
| content | one of the three | The content as text, up to 50,000 characters |
| title | no | The title, when passing content as text |
Results cover the pages crawled so far and name the crawl date. Pages published since the last crawl are not included.
Ask: "Plan internal links for the SBOM guide draft."
AI search visibility¶
visibility_check_topic¶
Answers "are we showing up in AI search for X" from the visibility results Hi, Moose has stored. Pass the question in your own words; Hi, Moose matches it to the tracked prompts, in any language.
| Input | Required | Meaning |
|---|---|---|
| question | yes | The topic or question, up to 400 characters |
| lookbackDays | no | How many days of stored results to read. Default 90. |
| engine | no | Only this engine, by name, such as "ChatGPT" |
Returns, per engine, whether the brand was mentioned or cited in the latest run and in an earlier run, the mention and citation rates over the window, a short excerpt from the latest answer where the brand appeared, the prompts matched, and other brands seen. Each excerpt carries an answer id for the next tool. Results cover the tracked prompts on the tracked engines only, never the whole of AI search. It reads stored results and never runs new checks.
Ask: "Are we showing up in AI search for SBOM generation?"
visibility_get_answer¶
Fetches one stored answer by its answer id: the prompt, engine, date, whether the brand was mentioned or cited, the sources the engine listed, the scorer's notes, and the full answer text. The text is what the engine wrote at the time.
Ask: "Show me the full ChatGPT answer from that result."
visibility_list_opportunities¶
Returns the ranked work queue Hi, Moose builds from stored visibility results, the same rows as the Opportunities tab. Each row names the exact gap, the one-sentence next step, the priority with its reasons, how consistently the gap held across runs, the effort, the status the team set, and an opportunity id.
| Input | Required | Meaning |
|---|---|---|
| lookbackDays | no | Days of stored results to read. Default 7, the same as the Opportunities tab. At most 90. |
| limit | no | Rows to return, best first. Default 10, at most 25. |
| action | no | Only one kind of row: create-coverage, improve-page, earn-inclusion, correct-description or investigate-change |
| priority | no | high, medium or low |
| includeClosed | no | Also return rows marked done, dismissed or snoozed. Default false. |
| opportunityId | no | An id from an earlier result. Returns the full evidence for that one row instead of the list. |
The five kinds of row: create-coverage (no page of yours answers a prompt competitors win), improve-page (a page of yours is cited or is the closest match but the brand is left out or ranked behind), earn-inclusion (a third-party page the engines cite names competitors but not you), correct-description (answers repeat a claim about the brand that needs checking) and investigate-change (the brand was named or cited before and is not any more).
Ask: "What should we work on for AI search this week?" then "Show me the evidence for the first one."
get_gap_analysis¶
Compares your brand's AI search visibility with one tracked competitor over a period, from the answers Hi, Moose has stored. It is the same report as the Gap analysis button on a competitor row under Visibility, Competition. It reads stored results and never runs new checks.
| Input | Required | Meaning |
|---|---|---|
| competitor | yes | The competitor's name or domain. It must be one of the project's tracked competitors. |
| period | no | A start and end date, from and to, as YYYY-MM-DD. Defaults to the last 7 days, the period the Visibility screen opens with. |
| project | no | Optional when your account has one project. |
Returns share of voice for both brands, how often each is named in answers, mention share and citation share, share of voice by engine, the topics where the competitor is named more, the sources cited in answers that name the competitor and not you, and a prioritised list of action items. Each action item has a priority, the action, the reason with its evidence, a type (create a page, improve a page, get mentioned, or protect), the topic, the engines and the sources. These are the same columns as the Excel and CSV export in the app. Where Hi, Moose has a tool for the work, the action item also names it: a content brief for a new page, a content refresh for a page to rework, or FAQs or key points to add to a page. The result ends with a link that opens the report in Hi, Moose.
If the competitor is not tracked, no report is built. The tool returns the tracked competitors so the app can ask which one you meant. The report is part of paid plans, including paid plans where you bring your own OpenRouter key. On a free plan no report is built and the tool says so, with a link to the plans.
Ask: "How do we compare with Globex in AI search?"
Visibility data for reports and dashboards¶
Three tools return the stored visibility data as structured JSON with a published output schema, so a Skill, a script or a dashboard can rely on the field names. They take the same filters as the Visibility screen in Hi, Moose. Every number covers the tracked prompts on the tracked engines on the days runs happened, never the whole of AI search.
visibility_list_tracked¶
Lists what the project tracks and the exact values the other two tools accept as filters: every tracked prompt with its id, topics, engines and whether it is branded or non-branded; the topics and engines with prompt counts; the tracked competitors; the brand name, aliases and owned domains; the run schedule; and the span of stored answers (first and last day, answers stored, engines and topics seen). The app calls this first when you name a topic, engine, prompt, page or competitor.
Ask: "What do we track in AI search for Acme?"
visibility_get_summary¶
Returns the numbers behind the Visibility dashboard for one date range.
| Input | Required | Meaning |
|---|---|---|
| from, to | no | The window as YYYY-MM-DD. Default: the last 30 days. At most 400 days. |
| lookbackDays | no | Window length when from is not given. Default 30. |
| topics, engines, prompts | no | Only these topics, engines or prompts, by the exact values from visibility_list_tracked. A plain engine name such as "ChatGPT" also matches its local-fetch variant. |
| promptType | no | branded (the prompt names the brand, an alias or an owned domain) or non-branded |
| pages | no | Only prompts whose answers cited one of these pages, then every engine for those prompts. Takes page URLs or bare domains such as example.com, your own or any other site |
| sentiment | no | positive, mixed, neutral or negative. Applies to answers where the brand was mentioned. |
| narrative | no | aligned, partial or drifted, the narrative verdict against the Brand Truth Profile |
| featureParity | no | accurate, needs-review, inaccurate or not-applicable |
| mentioned, cited | no | true or false, to keep only answers that mention or cite the brand, or only those that do not |
| competitor | no | Only answers that mention this competitor, by name |
| citationsList | no | For the citations block: both (default), domains or pages. Ask for one list alone to fit more rows in a result |
| citationsLimit | no | For the citations block: how many rows of each list to return per call, most cited first. Default 50, up to 150 |
| citationsOffset | no | For the citations block: skip this many domains and pages, to page through the full lists. Use the nextOffset of the previous call |
| minCitations | no | For the citations block: only domains and pages cited at least this many times in the window |
| citedDomains | no | For the citations block: exact counts for only these domains and the pages on them, however rarely they were cited |
| mergeEngineVariants | no | For the citations block: count an engine's local-fetch variant under the engine's plain name in the per-engine counts. Default true |
| sections | no | Which blocks to return: totals (always), engines, topics, prompts, trend (per day), competitors (share of voice, per engine and topic, pages competitors get cited from, prompts competitors win), citations (pages and domains the engines cite, own pages flagged), mentions (every brand named, with rank and sentiment). Default totals, engines and topics. |
| compare | no | true also returns the totals for the window of the same length immediately before |
| compareFrom, compareTo | no | A custom comparison window instead |
Each block carries the answer count, how many were unavailable, mentioned and cited counts, mention and cite rates, the average rank of the brand, and the sentiment, narrative and feature parity counts. Rates are fractions of answers; unavailable answers are left out of them.
Ask: "Give me the visibility summary for September with a comparison to August, by engine and topic." or "Share of voice against competitors for the SBOM topic, last 90 days."
visibility_list_results¶
Lists the stored answers one row per prompt, engine and run, newest first, with the same filters as the summary.
| Input | Required | Meaning |
|---|---|---|
| from, to, lookbackDays, topics, engines, prompts, promptType, pages, sentiment, narrative, featureParity, mentioned, cited, competitor | no | The same filters as visibility_get_summary |
| fields | no | compact (default) or full. Full adds the scorer reasons, the sources the engine listed, every brand named with its rank and sentiment, and a short excerpt of the answer. |
| sort | no | newest (default) or oldest |
| pageSize | no | Rows per page. Default 50, at most 200. |
| cursor | no | The nextCursor from the previous page, with the same filters |
Each row carries the answer id, date, run id, engine, topic, prompt, branded or non-branded, whether the brand was mentioned and cited, its rank, the page cited, the sentiment, narrative and feature parity verdicts, and the competitors mentioned and cited. The answer id works with visibility_get_answer for the full text.
Ask: "List every answer from the last two weeks where a competitor was named and we were not, as a table." or "Export the Perplexity results for October to a spreadsheet."
The output schema of each tool is part of the tool list the server publishes at https://mcp.himoose.com/tools.json, so you can hand it to the assistant that writes your Skill or script.
Audio¶
create_audio¶
Creates a "listen to this article" episode: a two-voice spoken conversation about the content, rendered as an MP3 and hosted by Hi, Moose, plus the HTML snippet for the embeddable player.
| Input | Required | Meaning |
|---|---|---|
| url | one of the three | A live page |
| draftId | one of the three | A saved draft |
| content and title | one of the three | The article as text, 1,200 to 200,000 characters, with its title |
| length | no | short (about 4 minutes), standard or deep. Default short. |
| hostVoice, guestVoice | no | Two different voices. Default Sulafat and Fenrir. |
| language | no | Narration language. Default: the language of the content. |
| focus | no | What the conversation should focus on, up to 500 characters |
| theme, layout, color, showTitle, customTitle, showTranscript, hideAttribution | no | The player look, see below |
The audio settings page lets you hear every voice, pick the length, language and player look, and copy a settings line to paste into the chat. While the episode renders the host device works on nothing else. If the result says the episode is still rendering, the next tool picks it up a few minutes later.
Ask: "Turn https://acme.example/guides/sbom into audio, short, with Zephyr and Charon."
get_audio_embed¶
Returns the player snippet for an existing episode, with a new look when player fields are passed. The audio is not re-rendered. Also the way to pick up an episode that was still rendering.
| Input | Required | Meaning |
|---|---|---|
| audioId | yes | The id from the episode |
| theme | no | dark, light or auto. Default dark. |
| layout | no | full or compact. Default full. |
| color | no | Accent colour as hex, for example #5688b0 |
| showTitle, customTitle | no | Show a title line, or set your own |
| showTranscript | no | Show the transcript button. Default true. |
| hideAttribution | no | Hide the "Audio by Hi, Moose" line. Paid plans only. |
Ask: "Give me the compact light player for that episode."
What the tools cannot do¶
They cannot change settings or team members, change what Hi, Moose tracks, start visibility runs, or delete anything. For dashboards and automation, use the Hi, Moose API.