Tools reference

The Mentionable MCP server exposes 12 tools on a single endpoint (POST https://mentionable.ai/api/mcp). 9 are read tools, 3 are write tools.

Each tool has its own dedicated page with full input schema, response shape and a real JSON example.

All tools

Tool	Purpose	Type
list_projects	List projects accessible to the API key	Read
list_prompts	Tracked prompts with mention rate per LLM	Read
list_fan_outs	Raw fan-out queries the LLMs run, deduped and ranked	Read
list_llm_sources	Domains cited or searched by LLMs	Read
list_competitors	Tracked competitors with mention totals and Share of Voice	Read
list_competitor_sources	Domains where a given competitor is cited	Read
list_backlink_opportunities	High-impact domains with marketplace offers and prices	Read
list_reddit_threads	Reddit threads cited by LLMs, with GEO signals and scraped content	Read
get_reddit_thread	Fetch a single Reddit thread by ID (poll enrichment status)	Read
bulk_update_competitor_status	Confirm or reject suggested competitors in bulk	Write
bulk_update_reddit_thread_status	Mark Reddit threads as commented, skipped or re-opened in bulk	Write
enrich_reddit_thread	Trigger Bright Data scrape of a Reddit thread (charges AI credits)	Write

Common conventions

• Pagination. Every list tool accepts cursor and limit (1 to 100, default 20). Responses include a pageInfo object with hasMore, nextCursor and totalCount.
• LLM enum. CHATGPT, CLAUDE, PERPLEXITY, GEMINI, GROK, COPILOT, GOOGLE_AI_MODE, GOOGLE_AIO.
• IDs. All id, projectId, promptId, competitorId are CUIDs.
• Dates. All timestamps are ISO 8601 UTC strings.
• Errors. Project not found, Project not in API key scope, Competitor not found, validation errors from Zod.

Auth requirements

Read tools require any non-customer role. Write tools require member or above. Every project-scoped tool checks the API key's effective scope before dispatch. See Authentication for details.