list_responses, when it asks whether to use it. Every tool carries a readable title as well; both are listed below.
Reading
List projects
List projects
list_projectsReadsThe projects (tracked brands) your account can see, with the workspace each belongs to, how often each one’s prompts run, and when its next scheduled run starts — or why none will. Assistants call it first to find the project you mean.You can narrow it down by
- Include archived projects
- The next run is the one the app’s top bar counts down to, in UTC, with the time left until it at the moment of asking.
- No run is planned for a paused or archived project, for a new project until its first run has succeeded, or for a workspace without a plan; the answer says which.
- Extra runs started by hand are not part of the schedule.
Get project overview
Get project overview
get_project_overviewReadsYour brand and every tracked competitor side by side: visibility, share of voice, position and sentiment, each with its change against the period before and whether that change is larger than the measurement’s margin.You can narrow it down by
- Time window: the last hour up to the last year, everything, or two dates of your own
- Compare with a period you name instead of the one just before, for example the week before a launch against the week after
- Only certain AI engines, topics, personas or countries
- Only the answers to one prompt, archived or not
- Break your own numbers down by AI engine, persona or topic in the same answer, with every competitor in each group and the strongest one named
- Which AI engine, persona or topic changed most in visibility, by the answers it lost or gained, with the measurement’s margin
- Measure every group against one competitor you choose
- How often you are named early, in the middle or late in an answer, and how many answers about you read positive, neutral or negative, as on the Overview page
- The early, middle and late shares and the positive, neutral and negative shares are for the chosen period only, without a change against before.
- Without sentiment in your plan, sentiment is left out, the positive, neutral and negative shares with it.
List prompts
List prompts
list_promptsReadsYour tracked prompts (the questions Palagus asks AI engines), each with its mentions, visibility, share of voice, position and sentiment, and how each moved.You can narrow it down by
- Time window: the last hour up to the last year, everything, or two dates of your own
- Only certain topics, personas, countries or AI engines
- Sort by any of the numbers or by the largest rise or fall
- Archived prompts instead of active ones
- Returns up to 200 prompts and says when there were more.
List competitors
List competitors
list_competitorsReadsThe competitors your project is measured against, with their websites and other names, how each was added, and whether any answer mentions it.Permission: Read your projects and metrics · Every role in the project, viewers included
Get top sources
Get top sources
get_top_sourcesReadsThe websites AI answers cite for your prompts, ranked, each marked as your own, a competitor’s or a third party’s, with two counts: every citation, and the number of answers that cite the site.You can narrow it down by
- Time window: the last hour up to the last year, everything, or two dates of your own
- Only your own, competitors’ or third-party sites
- Only certain AI engines, topics, personas or countries
- Only the answers to one prompt, archived or not
- Which kinds of third-party sites are cited (shops, media, review sites, …) and which kinds of pages (product pages, articles, forum threads, …), as in the Domain Categories and URL Types charts
- Up to 50 sites
- Needs the citations feature in your plan.
- Counts websites (domains). Which single pages an answer cited is in the answer itself (Get response).
- The kinds of sites and pages are counted for the chosen period only, without a change against before. Pages Palagus has not classified yet are counted apart.
List responses
List responses
list_responsesReadsThe AI answers collected for your prompts, one excerpt each: which engine answered, from which country, whether and how your brand came up, and which competitors were named.You can narrow it down by
- Time window: the last hour up to the last year, everything, or two dates of your own
- Search the answers’ own words
- Only answers that mention you, or only those that do not
- Only answers that name a given competitor, with or without you
- Only the answers to one prompt
- Only certain AI engines, countries, topics or personas
- Sort by date, sentiment, position or number of mentions
- Up to 50 answers per page, with as many pages as there are answers.
Get response
Get response
get_responseReadsOne AI answer in full: the complete text, the pages it cites, and how your brand came up in it.Permission: Read your projects and metrics · Every role in the project, viewers included
List personas
List personas
list_personasReadsThe personas your prompts are asked as (a buyer, a beginner, an expert), with how many prompts each one carries, so gaps in the prompt set show.Permission: Read your projects and metrics · Every role in the project, viewers included
Get project timeline
Get project timeline
get_project_timelineReadsHow visibility, position and sentiment developed over time for your brand and every competitor: one point per hour, day, week or month, each with the number of answers behind it.You can narrow it down by
- Time window: the last hour up to the last year, everything, or two dates of your own
- Day, week or month per point
- Only certain AI engines, countries, topics or personas
- Only one prompt
- Only some of the three curves
- Also the share of citations that point to your own website, as in the Domain Presence chart, with the number of citations behind each point
- The citation share is your own brand’s, not each competitor’s, and only when asked for. Without the citations feature in your plan it is left out with that reason; the other curves still come.
Search help
Search help
search_helpReadsSearches this documentation and a map of the Palagus app, so your assistant can answer how something works or where to find it, with a link to the page or the place in the app. Needs no project.You can narrow it down by
- How many pages to return
Read help page
Read help page
read_helpReadsReads one page of this documentation, or one place in the app with where it is and what it does, in full.Permission: Read your projects and metrics · Every role in the project, viewers included
List competitor suggestions
List competitor suggestions
list_competitor_suggestionsReadsThe brands AI answers name that your project does not track as competitors yet, as on the Suggested tab of the Competitors page: each with its website and other names, how many answers named it, and, where Palagus judges it is not a competitor, why (a reseller, a supplier, a media site, …).You can narrow it down by
- Up to 100 suggestions, in the tab’s order: the most-named first, those judged not a competitor last
- Counts every answer ever collected for the project, with no time window.
- Changes nothing: you add or dismiss a suggestion in the app, on the Suggested tab. The result carries that tab’s address, so your assistant can link it.
Adding and running
Create topics
Create topics
create_topicsCreatesAdds topics to a project to group prompts by theme. A topic that already exists is reported as existing, not created twice.You can narrow it down by
- Up to 50 topic names at once
- Counts against your plan’s topic quota, only names that are actually new.
Create prompts
Create prompts
create_promptsCreatesAdds prompts to a project and runs each one once, right away, on the AI engines you choose, so the first answers arrive without waiting for the schedule.You can narrow it down by
- Up to 50 prompts at once, each up to 250 characters
- Up to 10 AI engines this project may use
- A country the prompts are asked from
- A persona per prompt
- Up to 3 topics for all of them, created when missing
- Counts against your plan’s prompt quota.
Run prompt
Run prompt
run_promptStarts runsStarts one extra run of a prompt you already track, on one AI engine, on top of the project’s schedule, when a fresh answer is needed now.You can narrow it down by
- The AI engine
- A country
- Counts toward the workspace’s on-demand measurements for the month, a fair-use limit; the scheduled measurements are not affected.

