Run BotSeen from Claude, Cursor, and any AI assistant.
Connect BotSeen's MCP server once and your assistant can check your AI visibility, diagnose losses, run scans, and manage projects — in plain English, with your live data.
https://www.botseen.com/api/mcpClaude.ai connects over OAuth — no key to paste. 33 tools. Works everywhere MCP does.
Natural language, real actions
Just ask. Your assistant does the rest.
No dashboards, no exports. BotSeen becomes a set of tools your AI can call directly — on your live data.
“How visible is my brand in AI answers right now?”
list_projectsget_project_visibilityTwo-minute setup
Connect your client
Claude.ai and Desktop connect over OAuth — sign in and approve, no key. Claude Code and Cursor use your API key from Settings → Developer API.
- 1Open Settings → Connectors → Add custom connector.
- 2Paste https://www.botseen.com/api/mcp and click Connect.
- 3Sign in through BotSeen and approve the workspace. Done — no key to paste.
https://www.botseen.com/api/mcpOne connection
33 tools, from read-only to run-it-all
Everything the dashboard does. Reads are free; anything that spends credits says so, and assistants confirm with you first.
33 of 33 tools
list_projectsReadEvery project with its headline AI-visibility metrics.
get_project_visibilityReadShare of voice — you vs each competitor, with deltas since last scan.
get_project_setupReadA project's tracked prompts (with ids), competitors, and engines.
get_scan_resultsReadLatest answer per prompt × engine, with citations and wrong claims.
list_hallucinationsReadFalse claims AI engines made about your brands.
get_citation_mapReadDomains AI cites — where to get mentioned to improve visibility.
get_credit_balanceReadPlan, credit balance, and monthly allowance.
estimate_scan_costReadCredits a scan would cost before you run it.
get_scan_statusReadLive progress of a running scan.
list_schedulesReadAll automated scan schedules.
list_experimentsReadTracked experiment baselines.
get_diagnosisReadRead a completed Answer DNA diagnosis.
get_llms_txtReadRead the latest generated llms.txt.
create_projectProjectsCreate a project with prompts, competitors, and engines.
delete_projectProjectsPermanently delete a project and its data.
generate_buyer_promptsProjectsDraft buyer-intent questions for a site.
add_promptPromptsAdd a tracked buyer prompt.
update_promptPromptsEdit a tracked prompt.
delete_promptPromptsRemove a tracked prompt.
add_competitorCompetitorsAdd a competitor to track against.
delete_competitorCompetitorsRemove a competitor.
set_project_enginesEnginesChoose which AI engines a project scans.
create_scheduleSchedulesCreate a recurring automated scan (paid plans).
update_scheduleSchedulesChange a schedule's cadence, time, or engines.
set_schedule_enabledSchedulesTurn a schedule on or off.
delete_scheduleSchedulesDelete a schedule.
run_scanScansStart a full AI-visibility scan.
cancel_scanScansCancel an in-flight scan; refunds undelivered credits.
run_diagnosisAnswer DNADiagnose exactly why you're losing a prompt, with a fix plan.
run_simulationSimulatorPreview whether a content fix would win the citation.
track_experimentExperimentsSnapshot a prompt's visibility as a baseline.
delete_experimentExperimentsDelete a tracked experiment.
generate_llms_txtllms.txtCrawl the site and write its llms.txt.
Safe by design
OAuth or API key
Claude.ai uses OAuth — no key to paste. Standards-based PKCE flow.
Scoped to a workspace
A connection only ever sees the one workspace you approve.
Revoke anytime
Cut off a connection or rotate a key from Settings in one click.
Credit-safe
Credit-spending tools announce their cost so assistants confirm first.
Give your assistant AI-visibility superpowers.
Start free, generate an API key, and connect in two minutes.