Docs

Use Mentionly from your code and your AI assistant

Everything you see in the inbox is available through a read-only REST API and an MCP server, so you can pull mentions into your own tools or ask Claude or Cursor about them. New mentions can also be pushed to you by webhook or Slack.

API keys

Create a key in Settings → API keys in the app. Keys start with mk_live_ and are shown once. Send it as a bearer token:

Authorization: Bearer mk_live_…

A key can read every project you own or are a team member of.
Revoke a key any time in Settings; requests with it fail with 401.
The API works during your free trial and on paid plans. After a trial ends without a plan, requests return 402.

REST API

Base URL: https://app.mentionly.so/api/v1. Responses are JSON with a data array; errors return { "error": "…" }.

GET /projects

Your projects (brands): id, name, domain, website_url, created_at, last_scanned_at.

curl https://app.mentionly.so/api/v1/projects \
  -H "Authorization: Bearer mk_live_…"

GET /mentions

Mentions with their AI scores, newest first. Query parameters:

`project_id`	Limit to one project. Default: all projects you can access.
`status`	new, handled, dismissed or all (default all).
`min`	Minimum relevance score 0–5. 3+ means relevant.
`platform`	reddit, x, hackernews, stackoverflow, youtube, github, bluesky, devto, lobsters, forums, tiktok, news
`sentiment`	positive, neutral or negative.
`tag`	One AI tag (see Scores & tags).
`keyword`	A keyword id, to get mentions matched by that keyword.
`view`	leads (high buying intent) or cited (threads AI engines cite).
`since`	ISO date; only mentions found after it.
`limit`	1–200, default 50.

curl "https://app.mentionly.so/api/v1/mentions?min=3&view=leads&limit=20" \
  -H "Authorization: Bearer mk_live_…"

Each mention:

{
  "id": "…",
  "project_id": "…",
  "platform": "reddit",
  "subreddit": "SaaS",
  "author": "someone",
  "content": "Looking for an alternative to …",
  "matched_keyword": "your brand",
  "source_url": "https://www.reddit.com/r/SaaS/comments/…",
  "status": "new",
  "relevance_score": 5,
  "relevance_reasoning": "Asks for an alternative to a listed competitor.",
  "intent_score": 86,
  "sentiment": "neutral",
  "tags": ["buy_intent", "competitor_mention"],
  "is_promotional": false,
  "ai_cited_at": null,
  "mentioned_at": "2026-10-27T09:12:00Z",
  "created_at": "2026-10-27T09:30:41Z"
}

AI assistants (MCP)

Mentionly runs a read-only MCP server, so Claude Code, Claude Desktop, Cursor and other MCP clients can answer questions like “what are people saying about us this week?” or “which threads should I reply to?” from your live data. It uses the same API key.

Claude Code

claude mcp add --transport http mentionly https://app.mentionly.so/api/mcp \
  --header "Authorization: Bearer mk_live_…"

Cursor, Claude Desktop and other clients

{
  "mcpServers": {
    "mentionly": {
      "type": "http",
      "url": "https://app.mentionly.so/api/mcp",
      "headers": {
        "Authorization": "Bearer mk_live_…"
      }
    }
  }
}

The server authenticates with your API key in a request header. Connectors that only support sign-in (OAuth), such as ChatGPT's, can't connect yet.

Tools

list_projects: your projects (brands) and their ids.
mention_stats: Count mentions grouped by a dimension (platform, sentiment, AI tag, relevance level, keyword, day, week, author or status), with relevant/negative/lead counts per group and the change versus the previous period of the same length.
search_mentions: Find individual mentions (posts) with optional text search and filters. Returns content snippets, links, scores and tags. Use it to quote examples or answer 'what are people saying'.
keyword_report: List tracked keywords with type, context, platforms, paused state, 30-day volume and health (healthy / noisy = mostly irrelevant / quiet = no mentions).
ai_visibility: AI search visibility: how often ChatGPT, Perplexity and Gemini mention the brand for tracked questions (this period vs previous), average answer rank, and the community threads AI engines cite most.

Webhooks & Slack

In Settings → Notifications, add a Slack incoming webhook or any HTTPS URL. When a scan finds mentions at or above your relevance threshold, Mentionly sends them right away. Saved inbox views (feeds) can have their own alert channel.

Webhook payload (POST, JSON):

{
  "event": "mentions.found",
  "project": { "id": "…", "name": "Your brand" },
  "mentions": [
    {
      "content": "…",
      "source_url": "https://…",
      "platform": "hackernews",
      "relevance_score": 4,
      "intent_score": 72,
      "sentiment": "positive"
    }
  ],
  "sent_at": "2026-10-27T09:31:02Z"
}

Scores & tags

Relevance (1–5): 5 is about your business, 4 an actionable opportunity around a competitor or your category, 3 relevant discussion, 1–2 tangential or a different meaning of the word.
Buying intent (0–100): how likely the author is looking for a product like yours right now. 60+ counts as a lead. Bug reports, code changes, churn and promotional posts are capped at 20.
Sentiment: positive, neutral or negative toward you or your category.