Attribution

Attribution answers one question: where did each paying fan come from, and how much did they pay? TEASE takes every transaction on your OnlyFans account, resolves each paying fan to a single acquisition source, and rolls the money up into a leaderboard that always reconciles to your account net — no double counting, no money that quietly disappears.

The result is a revenue-by-source ranking you can drill from source → country → link → fan, split by money type (subscriptions, rebills, PPV, tips), tagged new-vs-returning, with trial→paid conversion and the fan-economics ratios (ARPPU, APC, APV, ARPNU) computed for you.

The mental model

Every paying fan lands in exactly one of three buckets. The buckets are disjoint and their revenue sums to your account total.

Enrich «organic» fans from their DMs

Many organic fans leave clues about where they came from — and much more — inside their DM history. The «Фаны» tab (under Раскладка / Sources) reads each fan's conversation locally and, with one BYO-key LLM pass per fan, extracts:

• a portrait — a 13-field sales dossier (who they are, what they buy, how to approach them);
• their country — inferred from what they tie to themselves (a place, a language, a timezone);
• their discovery source — when they reference it (a platform they mention browsing, or a content-format tell like "your reel" → Instagram).

Pick how many fans to process on the slider and press «Запустить разбор N фанов» — once. The panel then drives the whole sweep for you: the engine slowly and safely pulls each fan's DM history from OnlyFans (top-spenders first, one small batch per engine cycle), and we portrait each fan as their conversation lands. You never re-click — cards fill in over time, and the progress line shows queued / captured / portraited. «Остановить» halts further harvesting (batches already in flight finish on their own). A confident DM source (≥ 0.6) replaces that fan's organic bucket in the leaderboard; weaker signals still show on the fan card without touching the money split. Country is the richest DM signal — an organic fan rarely states their source outright, so treat source-from-DM as a bonus on top of the portrait + country.

The sweep is ban-safe by construction: the harvest is the engine's own passive, serial, human-paced capture (one thread at a time, self-halting on any throttle); the panel only queues the next batch — it never drives OnlyFans directly — and the portrait is one local LLM turn over already-captured text. Every read is per-creator scoped.

Read the leaderboard

The leaderboard is windowed by the range you pass and by the lifetime toggle (see Windowing). Read it from /api/admin/stats and take revenue.sources.

curl "https://app.tease.link/api/admin/stats?range=30d" \
  -H "Authorization: Bearer $TEASE_API_KEY"

const res = await fetch(
  'https://app.tease.link/api/admin/stats?range=30d',
  { headers: { Authorization: `Bearer ${process.env.TEASE_API_KEY}` } },
);
const stats = await res.json();
const leaderboard = stats.revenue.sources;

import os, requests

res = requests.get(
    "https://app.tease.link/api/admin/stats",
    headers={"Authorization": f"Bearer {os.environ['TEASE_API_KEY']}"},
    params={"range": "30d"},
)
leaderboard = res.json()["revenue"]["sources"]

Query parameters

Response — revenue.sources

{
  "sources": [
    {
      "domain_source": "ig-main",
      "source_label": "Main IG",
      "channel": "instagram",
      "subs": 312,
      "spenders": 188,
      "revenue_net_cents": 421550,
      "revenue_gross_cents": 502430,
      "pct": 0.41,
      "new_fans": 240,
      "returning_fans": 72,
      "trial_subs": 96,
      "trial_to_paid": 51,
      "funnel": { "new": 205, "old": 37, "ftl": 96, "paid": 112, "paid_after": 74, "ftl_paid": 38, "entered": 250 },
      "by_kind": [
        { "kind": "subscription", "net_cents": 184200, "count": 240 },
        { "kind": "rebill",       "net_cents": 96100,  "count": 130 },
        { "kind": "ppv",          "net_cents": 121050, "count": 410 },
        { "kind": "tips",         "net_cents": 20200,  "count": 88 }
      ],
      "countries": [
        { "country": "US", "subs": 140, "spenders": 92, "revenue_net_cents": 210400, "revenue_gross_cents": 248900 },
        { "country": "GB", "subs": 51,  "spenders": 33, "revenue_net_cents": 74100,  "revenue_gross_cents": 88200 }
      ],
      "links": [
        { "link_ref": "lnk_8fq2", "subs": 201, "spenders": 130, "revenue_net_cents": 290300, "revenue_gross_cents": 345100 }
      ]
    }
  ],
  "organic": {
    "domain_source": "__organic__", "channel": "", "subs": 410, "spenders": 260,
    "revenue_net_cents": 318900, "revenue_gross_cents": 379400, "pct": 0.31,
    "new_fans": 410, "returning_fans": 0, "trial_subs": 0, "trial_to_paid": 0,
    "funnel": { "new": 0, "old": 0, "ftl": 0, "paid": 0, "paid_after": 0, "ftl_paid": 0, "entered": 0 },
    "by_kind": [], "countries": [], "links": []
  },
  "pending": {
    "domain_source": "__pending__", "channel": "", "subs": 88, "spenders": 0,
    "revenue_net_cents": 0, "revenue_gross_cents": 0, "pct": 0.0,
    "new_fans": 88, "returning_fans": 0, "trial_subs": 88, "trial_to_paid": 0,
    "funnel": { "new": 0, "old": 0, "ftl": 0, "paid": 0, "paid_after": 0, "ftl_paid": 0, "entered": 0 },
    "by_kind": [], "countries": [], "links": []
  },
  "total_net_cents": 1025840,
  "total_gross_cents": 1221230,
  "by_kind": [
    { "kind": "subscription", "net_cents": 402100, "count": 590 },
    { "kind": "rebill",       "net_cents": 221400, "count": 318 },
    { "kind": "ppv",          "net_cents": 351900, "count": 1204 },
    { "kind": "tips",         "net_cents": 50440,  "count": 210 },
    { "kind": "other",        "net_cents": 0,      "count": 0 }
  ],
  "coverage": { "attributed_subs": 740, "organic_subs": 410, "pending_subs": 88, "trial_unknown_subs": 88 },
  "trial_subs": 184,
  "trial_to_paid": 97,
  "reconciliation": {
    "balanced": true,
    "sources_net_cents": 706940,
    "organic_net_cents": 318900,
    "pending_net_cents": 0,
    "total_net_cents": 1025840
  },
  "since": "2026-06-01"
}

Leaderboard fields

Source row fields

Each entry in sources (and the organic / pending rows) has this shape:

Full Attribution funnel

Every real-source, organic, and pending row always carries the same required funnel object. It counts entry-ledger events plus live re-subscribes for the requested window; calculating it never changes the row's money, lifecycle fields, or reconciliation. For a preset, or a custom upper bound equal or clamped to request time, this funnel has an open upper edge: its calculation samples current time and accepts entry/return timestamps strictly before that sample plus one hour. This is a vendor-clock-skew allowance, not an exact to_ts + 3600 boundary. A historical custom upper bound below request time stays exact.

new_fans and funnel.new answer different questions and can differ. new_fans is the leaderboard's lifecycle complement of returning_fans across all subs assigned to the row. funnel.new is window-scoped: it requires an entry inside this request's effective entry window and no transaction earlier than its one-hour lower-bound tolerance. Lifecycle new_fans uses the fixed account tracking anchor instead. Neither field is a subset of the other; when a window crosses that anchor, either count can be larger.

Money by type

by_kind appears at two levels — globally (revenue.sources.by_kind) and per source (row.by_kind) — and uses the same ordered keys every time:

Each entry is { "kind", "net_cents", "count" }. Within any bucket, Σ net_cents reconciles to that bucket's revenue_net_cents, so a "subscriptions vs PPV" chart never drifts from the headline.

Per-country drill

Every source row carries a countries array. Each entry breaks that source's revenue down by the fan's country:

{ "country": "US", "subs": 140, "spenders": 92, "revenue_net_cents": 210400, "revenue_gross_cents": 248900 }

Per-link drill

The links array on each source row attributes revenue down to the individual smart link the fan arrived through:

{ "link_ref": "lnk_8fq2", "subs": 201, "spenders": 130, "revenue_net_cents": 290300, "revenue_gross_cents": 345100 }

For a geo-by-link matrix across a domain's links, read the dedicated breakdown:

curl "https://app.tease.link/api/admin/domain_links/smartlinks/breakdown?domain=yourbrand.link" \
  -H "Authorization: Bearer $TEASE_API_KEY"

const res = await fetch(
  'https://app.tease.link/api/admin/domain_links/smartlinks/breakdown?domain=yourbrand.link',
  { headers: { Authorization: `Bearer ${process.env.TEASE_API_KEY}` } },
);
const breakdown = await res.json();

res = requests.get(
    "https://app.tease.link/api/admin/domain_links/smartlinks/breakdown",
    headers={"Authorization": f"Bearer {os.environ['TEASE_API_KEY']}"},
    params={"domain": "yourbrand.link"},
)
breakdown = res.json()

This returns each link with its per-country and per-source split, so you can compare which link on a domain pulls which geographies. The money reconciles to the same per-link totals you see in the leaderboard's links array.

Per-fan drill

Three reads take you from a source row down to the individual fans behind it.

curl "https://app.tease.link/api/admin/sources/yourbrand.link/spenders" \
  -H "Authorization: Bearer $TEASE_API_KEY"

curl "https://app.tease.link/api/admin/spenders?range=30d" \
  -H "Authorization: Bearer $TEASE_API_KEY"

curl "https://app.tease.link/api/admin/fan/FAN_ID" \
  -H "Authorization: Bearer $TEASE_API_KEY"

New vs returning, and trial → paid

Two product questions are answered inline on every row, so you don't have to compute them.

• New vs returning — the money-history branch uses the fixed account funnel_tracking_since anchor, not the selected money window. returning_fans includes fans with an earlier transaction and windowed live re-subscribes whose current of_fans action is return; the rest are new_fans. The aggregate row is the authoritative split: legacy per-fan drill flags only reflect the transaction-history branch.
• Trial → paid — trial_subs is the count of free-trial fans in the bucket; trial_to_paid is how many of them later produced paid revenue. Read together they are your trial-funnel conversion for that source.

Coverage census

coverage is an honest sub-level census — how many fans landed in each bucket — so you can see attribution completeness at a glance. It moves no money, so it never affects reconciliation.

Fan economics

The economics ratios are returned as outputs in two scopes alongside the leaderboard: revenue.fans (attributed scope) and revenue.fans_account (whole-account scope). You read them; TEASE computes them.

{
  "scope": "account",
  "arppu_cents": 2240,
  "apc": 3.1,
  "apv_cents": 720,
  "arpnu_cents": 1180,
  "paying_fans": 458,
  "payments": 1420,
  "new_fans": 240,
  "new_paying_fans": 132,
  "new_to_paying_pct": 55.0,
  "cohort_reliable": true,
  "cohort_revenue_cents": 283200,
  "maturing": false
}

Windowing: lifetime vs selected range

Attribution supports two windowing modes, chosen with the lifetime parameter:

• Selected range (default, lifetime=false) — the «new traffic» view. Totals use the exact range or explicit from_ts/to_ts; the account tracking anchor does not floor money. The since field is a legacy anchor label, not the effective lower bound.
• Lifetime (lifetime=true) — the «за всё время (LTV)» view. Without a complete custom from_ts/to_ts pair, totals and grouped aggregates use every allowed revenue transaction all-time and ignore range; since reads "lifetime". Day-based trend axes are intentionally limited to the latest 730 days and coarsened for long spans. A complete custom pair overrides lifetime.

Both modes preserve the reconciliation invariant — the buckets always add up to the total for whichever window you asked for.

Reconciliation

The reconciliation object is your integrity check on the leaderboard:

{
  "balanced": true,
  "sources_net_cents": 706940,
  "organic_net_cents": 318900,
  "pending_net_cents": 0,
  "total_net_cents": 1025840
}

FAQ

Next steps

• Smart Links — create the trackable links these sources resolve to.
• Telegram attribution — the same question for the Telegram sales bot's revenue.
• Webhooks & CAPI — forward attributed conversions to Meta and TikTok.
• API Reference — the full /api/admin/stats schema.