LLM Mentions API

Get AI Visibility DataAcross ChatGPT, Claude, Gemini & Perplexity with one API call

Query 4 LLM APIs (ChatGPT, Claude, Gemini, Perplexity) in parallel. Plus real-time Perplexity UI scraping for ground truth. One bearer token, one response shape.

300 free credits · No card required · Subscribe when you need more

# Query 4 answer engines in one call
curl https://api.mentionsapi.com/v1/ask \
  -H "Authorization: Bearer $MENTIONSAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "providers": ["openai", "anthropic", "gemini", "perplexity"],
    "prompt": "What are the best project management tools in 2026?",
    "track_brands": ["Notion", "Linear", "Asana"]
  }'
Credit-based·300 free credits · $5/mo → 1,000 credits · $54/yr → 12,000
Why MentionsAPI

Everything you need to ship an AI-visibility feature.

Each one is the thing you'd normally have to write yourself. We did it once, so you don't.

One API, four answer engines

Query GPT-5, Claude 4, Gemini 2.5, and Perplexity Sonar through a single endpoint. One bearer token, one rate limit, one response shape. Swap providers by editing an array.

providers: ["openai", "anthropic",
            "gemini", "perplexity"]

Brand tracking built in

Pass a track_brands array and get back position, sentiment, surrounding context, and every cited URL per mention. No regex, no custom parser, just a clean array your dashboard can render.

{ brand: "Notion",
  rank: 1,
  sentiment: "positive" }

Cached and rate-limit-safe

Instant low-cost answers for recently-asked queries, automatic retries, per-key rate limits, and partial-success fallbacks when a provider degrades. Your monitoring dashboard never takes down a provider outage with it.

// automatic retries + cache
x-cache: "HIT"
x-cache-tier: "shared"
How it works

From zero to structured brand data in three steps.

No SDK to install. No parser to write. A bearer token and a JSON body is the whole integration.

  1. STEP 01

    Authenticate with a bearer token

    Create an API key in the dashboard. One key, one Authorization header. Good across every provider we support.

     bash
    curl https://api.mentionsapi.com/v1/ask \
      -H "Authorization: Bearer lvk_live_xxxxxxxxxxxxxxxx"
  2. STEP 02

    POST a prompt with providers and brands

    One request fans out to every provider you list, in parallel. Pass track_brands to get structured mentions back.

     POST /v1/ask
    {
      "providers": ["openai", "anthropic", "gemini", "perplexity"],
      "prompt": "What are the top CRMs for startups?",
      "track_brands": ["HubSpot", "Salesforce", "Attio"]
    }
  3. STEP 03

    Read structured mentions and citations

    Every provider returns the same shape. Loop through results; render mentions by rank, sentiment, or competitor cohort.

     JavaScript
    for (const m of data.brand_mentions) {
      console.log(m.provider, m.brand, "#" + m.rank, m.sentiment);
    }
    // openai    HubSpot #1 positive
    // openai    Attio   #2 positive
    // anthropic Attio   #1 positive
Who it's for

Built for teams shipping AI-visibility tools in 2026.

Four common shapes. Same API underneath. Start free, upgrade when your product ships.

Built for the builders shipping AI-visibility tools
Indie developersSEO agenciesGEO tool buildersBrand monitoring developersContent marketersData engineers
Pricing

Simple plans. Credits do the metering.

Subscribe from $5/month and every call debits credits by what it actually runs. Repeat checks cost a handful of credits; premium 4-provider fan-outs are 250.

Per-call pricing
Quick mode
/v1/check?mode=quick. 4 LLM APIs in parallel, fastest path.
130 credits
5 when cached
Perplexity Live
/v1/check?mode=perplexity_live. UI scrape with full citations + fan_out.
85 credits
per call
Discover
/v1/discover. 10 query candidates for any brand.
165 credits
per call
Compare
/v1/compare. Head-to-head delta between two brands.
500 credits
per call
From early users

The shape of feedback we keep hearing.

We're in public beta. Quotes below are paraphrased from early users with their details anonymized on request. Email us to talk to a reference.

We dropped three SDKs and a normalization layer in the same PR we rolled MentionsAPI out. The biggest surprise was how accurate the sentiment tagging is. We stopped second-guessing it after the first week.
Engineering lead
Growth-stage B2B SaaS
We were polling four providers daily for 20 clients. After we switched, cache hits did what we hoped: same freshness, roughly a quarter of the bill. Our client dashboards didn't notice a difference.
SEO agency lead
Marketing agency, EU
I wired an internal brand-watch tool in a weekend. The structured mention object (position, sentiment, cited URL) is the specific shape I'd have built myself after getting tired of regex.
Senior developer
Consumer marketplace
FAQ

Frequently asked questions

Full details on the pricing and methodology pages. Anything else, email [email protected] — we answer fast.

An API that checks whether a brand shows up in AI answers. One call queries the surfaces you pick and returns structured mentions — position, sentiment, surrounding context, and every cited URL — instead of raw text you'd have to parse yourself. One bearer token, one response shape, JSON in, JSON out.

Ship your AI-visibility feature this week.

Sign up free with 300 credits and make your first call in under 60 seconds. Subscribe for $5/mo when you need more.