Henneth for Claude and ChatGPT.
Ask Claude or ChatGPT about your own brands in Henneth. The connector reads your data after you sign in and shows nothing you could not already see in the app.
What the connector does
It reads the Henneth brands of the person who signs in, and makes two writes: asking an AI engine a question live, and creating a task.
It can read·Your brands and your role on each, plus the topics, tags and personas set up for them.·Visibility by answer engine, with change over time and rank.·How your brand compares with its competitors, overall, by topic or by persona.·Tracked prompts, the latest answer from each engine, sentiment and the searches engines ran.·The websites AI answers cite and the topics where citations go to competitors.·Ads shown in AI answers and the performance of your own ChatGPT ads.·Site readiness audits and which AI agents read your website.·Alerts, Stock Guard findings, product catalog issues and recommended actions.·Differences, gaps and rankings calculated from figures already returned.
It can write·Send a question to an AI engine live and return its answer and cited sources. This spends paid credits.·Create a task in your Action Center, only after you confirm.
Set up in Claude
Adding a custom connector may need an owner or admin on Team and Enterprise plans.
Steps01Open Settings, then Connectors.02Choose Add custom connector.03Enter the URL https://mcp.henneth.ai/mcp and choose Add.04Choose Connect.05Sign in with your app.henneth.ai account.06Choose Allow.
https://mcp.henneth.ai/mcpOnce Henneth is listed in the Connectors Directory, you can add it from there instead of typing the URL.
Set up in ChatGPT
Developer mode is available on some ChatGPT plans. Workspace admins may need to allow it first.
Steps01Open Settings, then Connectors, then Advanced, and turn on Developer mode.02Choose Create.03Enter a name, the URL https://mcp.henneth.ai/mcp, and set authentication to OAuth.04Choose Create.05Sign in with your app.henneth.ai account.06Choose Allow.
Once Henneth is in the Apps directory, enable it from the directory and skip Developer mode.
Tools
49 tools: 47 read and 2 write. Every tool except list_brands takes a brand from list_brands. Dates default to the last 28 days. The ads and brand-context tools answer with a plain "not enabled" note when the brand's workspace has not turned those features on.
| Tool | What it does | Access |
|---|---|---|
| Brands and scope | ||
list_brands | Lists the brands in Henneth that the signed-in user is a member of, with their role on each. | Read |
list_scope | Lists the topics, tags, personas and monitored AI engines set up for a brand. | Read |
get_quota | Shows a brand's monthly playbook run usage, what is left, the limit and the current period. | Read |
get_brand_context | Reads the company, audience, offerings, competitor, voice and strategy notes the brand's team has written down, or says plainly that the feature is not enabled. | Read |
| Visibility and competitors | ||
get_visibility_summary | Returns how often a brand appears in answers from each AI engine over a date range, with its change, rank and the top brands alongside it. | Read |
get_visibility | Returns a brand's headline visibility score for a date range, its change against the previous range and the daily trend, optionally for one topic, tag, persona or engine. | Read |
get_overview | Returns the headline visibility score, its change and the daily trend for a date range. | Read |
get_platforms | Compares and ranks all monitored AI engines by visibility and number of analyzed answers. | Read |
get_competitor_comparison | Ranks a brand against its tracked competitors by visibility over one date range, with each one's change against the range before. | Read |
compare_competitors | Compares a brand with its competitors in AI answers, overall or for one tracked topic. | Read |
get_topic_comparison | Compares a brand with its tracked competitors topic by topic, with one visibility figure per brand per topic. | Read |
get_persona_comparison | Compares the personas set up for a brand by their headline visibility over a date range. | Read |
get_absence_competitors | Names the tracked competitors that appear in the answers where the brand was not mentioned at all, and how many such answers each one fills. | Read |
get_branded | Lists the prompts that name brands, such as a brand against a rival, with how often the answers name the brand, its sentiment and which competitors appear. | Read |
| Prompts and answers | ||
list_tracked_prompts | Lists the prompts Henneth tracks for a brand, with the share of answers that mention the brand. | Read |
list_prompts | Lists every tracked prompt with its topic, tags, whether it is active or paused, when it was added and whether results exist yet. | Read |
get_prompts | Shows how tracked prompts perform, with topic, visibility over the date range and average mention position, for prompts that already have results. | Read |
get_prompt_detail | Reads one tracked prompt in depth, with the domains and pages its answers cite, how many answers name the brand and the newest answers. | Read |
get_prompt_answers | Returns the latest answer from each AI engine to one tracked prompt, with whether the brand and its competitors were mentioned and the sources cited. | Read |
get_answer | Reads one stored AI answer in full, with the engine, date, whether it names the brand, its sentiment, the competitors it names and the sources it cites. | Read |
get_query_fanouts | Lists the web searches AI engines ran while answering the tracked prompts, how often each ran and on which engines. | Read |
get_sentiment | Returns the positive, neutral and negative shares of the answers and the daily trend of the positive share. | Read |
get_sentiment_answers | Lists the stored answers behind one sentiment group, most extreme first, each with its engine, prompt, date and an excerpt. | Read |
| Citations | ||
get_citations | Lists the websites AI answers cite, with each one's share of citations, its change and its category, such as owned, competitor or earned media. | Read |
get_citation_domain | Reads one cited domain in depth, with its share of citations, the pages that get cited and which engines cite it. | Read |
get_citation_losses | Names the individual websites that decide the category, how often each is cited and whether it mentions the brand. | Read |
get_topic_citation_gaps | Names, for each topic the brand is losing, the websites cited in the answers it lost. | Read |
| Answer ads and your ads | ||
get_answer_ads | Names the brands whose ads were served in AI answers to the brand's prompts, as a share of answers on each platform, against the same number of days before. | Read |
get_ads_connection | Checks whether the brand's OpenAI Ads account is connected to Henneth and shows its name, currency and review status. | Read |
get_ads_campaigns | Lists the brand's ChatGPT ads campaigns for a period, with status, budget, impressions, clicks, click-through rate, cost per click and spend. | Read |
get_ads_performance | Shows how the brand's ChatGPT ads performed over a period, with the daily series, the worst campaign and the sessions its own analytics recorded from the same ads. | Read |
get_ads_history | Lists the changes made to the brand's ChatGPT ads in the last days, newest first, with who made each one and the outcome. | Read |
get_ads_gaps | Lists the tracked questions whose answers named a competitor and never the brand, most lost first, which are the gaps an ads campaign can answer. | Read |
get_ads_landing_pages | Lists pages of the brand's own site, read from its sitemap, that an ads campaign on a topic could send people to. | Read |
get_ga4_paid | Shows the sessions, engaged sessions and key events the brand's own Google Analytics recorded from its ChatGPT ads, split by ad. | Read |
| Site readiness and agent traffic | ||
get_audit | Returns the latest site readiness audit, with the score, the six dimension scores, findings by severity and the priorities, against the run before. | Read |
get_audit_history | Lists the completed readiness audits, newest first, with each score, its change and its open and critical findings. | Read |
get_audit_finding | Reads one finding of the latest readiness audit, with why it matters, the projected score lift, the effort and when it counts as done. | Read |
get_agent_analytics_overview | Shows which AI agents read the brand's own website, with visits, unique bots, pages crawled, the top pages and crawl health. | Read |
get_agent_analytics_pages | Ranks the pages of the brand's own website that AI agents read, by visits, with which platforms read each one. | Read |
get_agent_analytics_bots | Lists every bot that visited the brand's own website, with its platform, what its traffic is for, its visits and the pages it read most. | Read |
| Alerts, catalog and actions | ||
list_playbook_alerts | Lists recent alerts for a brand and the latest Stock Guard findings for products that AI answers recommend. | Read |
get_catalog_issues | Lists sync failures, stale syncs and product data gaps found in a brand's connected product catalogs. | Read |
get_actions | Lists the recommended and tracked actions for a brand, including quick wins and their stored impact scores. | Read |
| Calculations on stored results | ||
derive_difference | Subtracts one stored figure from another and returns a signed difference that can be cited. | Read |
derive_gap | Returns how far apart two stored figures are, with the direction from the brand's point of view. | Read |
rank_results | Orders the rows of a stored result by a chosen field and returns the top rows. | Read |
| Live query | ||
ask_ai_engine | Sends a question to the chosen AI engine live, returns that engine's answer and the sources it cited, spends paid credits and is limited to 20 calls per workspace per day. | Write |
| Tasks | ||
create_playbook_task | Creates one task in the brand's Action Center only after the user chooses a priority and explicitly confirms; a first call returns a proposal and writes nothing. | Write |
Authentication and permissions
·Sign-in is OAuth 2.1 with PKCE. You sign in with your Henneth account at clerk.henneth.ai, never with a password shared with the assistant.·The assistant receives a token that can only see brands you are a member of. Your Henneth role still applies to everything it does.·47 tools only read. Two write: ask_ai_engine asks an AI engine a question live and spends paid credits, and create_playbook_task creates a task only after you confirm. Neither can delete or change anything that exists.·Henneth receives the tool inputs the assistant sends, such as a brand and a date range. We never receive your chat history.·To revoke access, disconnect the connector in Claude or ChatGPT, or write to [email protected] and we will revoke it from the Henneth side.
Machine-readable details are in the MCP server card.
Example prompts
Which brands do I have in Henneth, and what is my role on each?
How did my brand's ChatGPT visibility change in the last four weeks?
Which competitors appear ahead of my brand in answers about my main topic?
Show the tracked prompts where my brand is mentioned least.
What did Gemini and Perplexity answer to my top tracked prompt yesterday?
Are there new Stock Guard findings or catalog issues I should fix this week?
Support
Write to [email protected] if the connector will not connect or answers something that looks wrong.Privacy policy →