Leaderboards

Leaderboards aggregate raw mentions or citations within a window. They answer "who's winning?" type questions: who's getting mentioned, which prompts surface them, where they're being cited.

All leaderboards share the same response envelope: { items: [...], from, to, truncated }. If truncated: true, the per-leaderboard scan cap was hit - narrow the time range for accurate counts.

The window is controlled by from and to (ISO 8601 date or datetime, inclusive). Both default to today (00:00 UTC → now). Maximum range is 30 days.

Top brands

In a project

Brands ranked by mention count within one analysis configuration's runs.

curl -H "Authorization: Bearer $KEY" \
  "https://api.42a.ai/api/v1/projects/prj_xxx/top-brands?from=2026-05-01&to=2026-05-14&limit=10"

const res = await fetch(
  `https://api.42a.ai/api/v1/projects/${projectId}/top-brands?from=2026-05-01&to=2026-05-14&limit=10`,
  { headers: { Authorization: `Bearer ${KEY}` } },
).then((r) => r.json());

for (const b of res.data.items) {
  console.log(`${b.name}: ${b.mention_count} mentions`);
}

import requests

res = requests.get(
    f"https://api.42a.ai/api/v1/projects/{project_id}/top-brands",
    params={"from": "2026-05-01", "to": "2026-05-14", "limit": 10},
    headers={"Authorization": f"Bearer {KEY}"},
).json()

for b in res["data"]["items"]:
    print(f"{b['name']}: {b['mention_count']} mentions")

Through a prompt

Brands ranked for a single prompt template. Useful for "for the prompt 'best CRM for early-stage SaaS', who is being recommended right now?".

curl -H "Authorization: Bearer $KEY" \
  "https://api.42a.ai/api/v1/prompts/prm_xxx/top-brands?from=2026-05-01&to=2026-05-14&limit=10"

Both endpoints return { brand_id, name, mention_count } per item.

Top prompts in a project

Which prompts in this project are pulling weight? Knowing this tells you what kind of content / SEO bets are paying off.

curl -H "Authorization: Bearer $KEY" \
  "https://api.42a.ai/api/v1/projects/prj_xxx/top-prompts?from=2026-05-01&to=2026-05-14&limit=10"

Each item includes prompt_template_id, name, category, version, is_active, and mention_count - enough to act on without a follow-up get_prompt call.

Top cited websites

When an LLM cites a source URL alongside a brand mention, that citation lives in the citations table. Each website row includes a domain, apex_domain, and a type classifier (news_article, review_site, academic, documentation, …).

In a project

curl -H "Authorization: Bearer $KEY" \
  "https://api.42a.ai/api/v1/projects/prj_xxx/top-websites?from=2026-05-08&to=2026-05-14"

In a vertical (prompt category)

A "vertical" is a prompt-template category (awareness, consideration, comparison, technical). This aggregates citations across every prompt in that category for the org - handy for "which sources do I need to win in 'comparison' content?".

curl -H "Authorization: Bearer $KEY" \
  "https://api.42a.ai/api/v1/categories/comparison/top-websites?from=2026-05-08&to=2026-05-14"

Truncation

Each leaderboard caps how many rows it will scan to keep responses fast:

When the cap is hit, truncated: true in the response and the counts are a strict lower bound. The fix is almost always to narrow the time range.

Common questions this answers

• "Who are my top X competitors in this project?" → top-brands in project.
• "Which prompts surface my brand the most?" → top-prompts in project.
• "Which sites should I get featured on for [vertical]?" → top-websites in vertical.
• "Where did the new mentions come from this week?" → top-websites in project with from set 7 days back.

See also

• Reference: project leaderboards
• Reference: prompt leaderboard
• Reference: vertical leaderboard

Last updated May 14, 2026