AB6 — Trending Audio Discovery System

You said your competitors have proprietary software you don't — software that scrubs TikTok, Instagram, and YouTube Shorts daily for new trending audio, filters out artists already signed to majors, runs an authenticity check, and surfaces unsigned candidates. This proposal is for that system, plus a structural improvement we want to flag before you decide.

The structural improvement

AB6's actual scene — cloud-rap, phonk, underground hip-hop — doesn't break TikTok-first the way mainstream pop does. The Suicide Boys / Pouya / Ghostemane lineage broke through SoundCloud, producer networks, festival circuit, and tour relationships, then hit TikTok months later. A pure trending-audio system catches the second wave, not the first.

So we structured this as a layered system with three pricing tiers. The base layer is the trending-audio system you described. The expansion layer adds scene-specific signal sources tuned to how AB6's roster actually broke through. You pick your tier.

Built specifically for AB6 — not a generic music-discovery tool. Every claim about data coverage was tested against your actual roster (Pouya, Fat Nick, Lu, KXLLSWXTCH, South Strip, Gutterboysouz, Mikey The Magician) before scoping. Results disclosed below — including the ones that didn't work.

Realistic timeline: candidates surfaced in Q1, conversations in Q2, signings in Q2-Q3, plaques in 2026 H2. The system surfaces signal; your A&R closes.

All three pricing tiers share the same backend: scrapers, scoring, filters, enrichment, persistence. The difference is which signal layers run on top, and which interface your team works in.

The system runs a parallel Spotify + Deezer lookup against every TikTok or Instagram trending sound and classifies the candidate into one of five distribution-status types, each with its own priority and handling logic:

• Type 1 · Pre-distribution. Spotify miss + Deezer miss. Artist active on TikTok / IG but hasn't released to DSPs — the earliest possible signal. Surface with prominent badge, run authenticity check, flag for immediate manual review. Highest priority.
• Type 2 · Standard candidate. Both Spotify and Deezer have the track. Run full ladder: ISRC registrant lookup → Deezer label field → Spotify popularity gate → Perplexity research if uncertain. Surface qualifying candidates with full metadata.
• Type 3 · Spotify-only. Mainstream-distributed but Deezer ingestion gap. Skip Deezer in the ladder; use Spotify popularity + ISRC fallback.
• Type 4 · Deezer-only. Unusual edge case (small US distributor missing from Spotify). Use Deezer label + ISRC. Flag for manual review since it's atypical.
• Type 5 · Roster-internal. Match via Deezer label string "All But 6 Records" / "AB6 Records" OR ISRC registrant prefix USUYG OR explicit name match. Auto-excluded.

Major-label filter — five-stage ladder

Spotify removed their structured album.label field in February 2026, so we recover that signal via a layered ladder. Tested every layer against your roster as part of scoping; results disclosed below.

• Stage 1 · ISRC registrant prefix lookup. Direct match against major-label registrant prefix table (Universal, Sony, Warner subsidiaries) → drop as signed. Includes USUYG for AB6 / EMPIRE roster exclusion.
• Stage 2 · Deezer label field. Free public API exposes record label per album. Parse {label} / {distributor} format and match against major-label patterns.
• Stage 3 · Spotify popularity gate. Popularity > 75 = drop as almost certainly signed. Popularity < 30 with low followers = almost certainly unsigned.
• Stage 4 · Perplexity Sonar research. Fires for the ~5-15% of cases stages 1-3 disagree on. Searches recent music industry trades, returns a verdict with cited source URLs. ~$5-15/mo at expected volume.
• Stage 5 · Manual override. Always available on the dashboard. Your team has the final say.

Each candidate displays the verdict, the reasoning, and (when applicable) a clickable source URL your team can verify before reaching out.

What we tested before scoping

• Deezer roster coverage: 6 of 6 tested AB6 artists found with full label data and ISRC codes. Even very small artists (Gutterboysouz at ~200 fans, South Strip at ~150 fans). Every release showed consistently as "All But 6 Records / EMPIRE" with ISRCs sharing the USUYG prefix — which is what makes the structural denylist work.
• Deezer major-label coverage: verified accurate label data on Drake (OVO Sound), Travis Scott (Cactus Jack / Epic), Carti (AWGE/Interscope Records), Future (Wilburn Holding Co./Epic), Bad Bunny (Rimas Entertainment LLC.).
• Deezer emerging-artist coverage: confirmed for Nettspend (10K fans), BLP Kosher (2.5K fans), OsamaSon (16K fans), 9mice (185 fans), LAZER DIM 700, DVRST, Kordhell, DJ Yung Vamp. Even artists with 185 fans had full catalogs indexed.
• MusicBrainz coverage: 1 of 7 hit rate. Cloud-rap and phonk artists are systematically under-represented in MusicBrainz's community-curated dataset. Kept in the ladder as a free supplementary check, but not load-bearing for your scene. Disclosed honestly.
• ISRC reverse-lookup: verified working via Deezer's /track/isrc:{code} endpoint. Note: ISRC codes can occasionally collide between original tracks and cover versions, so the system uses ISRC + title + artist normalization as a triple-check rather than ISRC alone. Drake's “God's Plan” ISRC GBDMT1800186 resolves to a metalcore cover by “Our Last Night” on Deezer — real edge case, mitigation built in.

Authenticity scoring — the “30 videos in Pakistan” filter

Multi-signal scoring (deterministic math, no AI judgment in v1 — math is more explainable, easier to override case-by-case): audience geography skew, post velocity profile, engagement-ratio anomalies, poster cluster analysis, account-age clustering. Each candidate gets a 0-100 score with 2-3 specific reasons. Advisory, not blocking. Your team makes the final call.

We are precise about these because the unofficial-endpoint reality is real and we'd rather you know the trade-offs upfront than discover them later.

• TikTok Creative Center is an unofficial endpoint. The Apify actor intercepts TikTok's own internal trends API. TikTok can change that internal structure. Apify usually patches within 1-3 days; we re-integrate within the maintenance retainer window. Historical breakage frequency: roughly once every 3-6 months.
• Instagram is materially more fragile than TikTok. Meta rotates internal endpoint identifiers every 2-4 weeks. Realistic expectation: 3-5 multi-day Instagram signal outages per year — about 15-25 broken-IG-signal days. Mitigation tiered: cookie refresh (auto), discovery account replacement (3-5 day gap), actor swap to one of 5 pre-vetted backups (3-7 day integration). All covered by the retainer. Plan for a fallback A&R cadence on those days.
• Major-label filter is heuristic, not a database lookup. Our 5-stage ladder achieves ~85-90% accuracy. Edge cases surface with verdict + reasoning + citation; manual override on the dashboard. Phase 2 upgrade to Soundcharts or Chartmetric improves to ~95% at $250-350/mo recurring.
• Spotify monthly listeners is scraped, not API'd. Spotify's official Web API doesn't expose monthly listener counts; we use an Apify scraper. If Spotify changes their HTML, the scraper breaks until updated. Optional feature — you can disable and rely on popularity score + follower count alone.
• Authenticity scoring is advisory, not blocking. 0-100 score with reasons. Your team makes the final call. Sophisticated paid promo (US-based real-account engagement farms) won't trip the heuristic — we catch dumb cases reliably and sophisticated cases imperfectly.
• Pre-distribution detection is the highest-value signal and the noisiest. True pre-release artists are rare; the dual-miss signal also fires on TikTok metadata bugs, regional-only releases, and beat-only audio. Real signal-to-noise gets validated and tuned during the week 4 tuning window.
• YouTube Shorts excluded from Phase 1 for structural reasons (no audio attribution on original creator uploads), not just cost. Reconstructing the signal would push recurring from $95-130/mo to $300-500/mo for a platform that lags TikTok in rap discovery anyway. Quoted as a tailored Phase 2 if signal proves missing after 60-90 days.
• Cloud-rap and phonk scenes don't break TikTok-first. The structural reason we recommend the Strategic tier. Layer 1 alone may catch the second wave instead of the first. If that happens with Essential or Recommended, the Layer 2 upgrade conversation becomes urgent.

Discovery accounts and brand protection

The system runs against TikTok and Instagram via dedicated discovery accounts that are completely isolated from AB6's real social media presence. We provision these, condition them (5-7 days of human-paced rap/phonk engagement), and operate them entirely on our side — your team never has to touch them. By keeping discovery accounts on Node-managed identities and Node-managed phone numbers, all unofficial-endpoint risk stays contained to throwaway infrastructure with zero exposure to AB6's real accounts. Your label's actual TikTok and Instagram presence remains untouched no matter what happens on the discovery side.

The Empire JV nuance

AB6's record label is a joint venture with Empire. Empire-distributed artists are a strategic gray zone — distribution-only deals don't preclude AB6 signing them, but ownership-stake deals do. The system handles this concretely:

• "All But 6 Records / EMPIRE" or "AB6 Records / EMPIRE" → AB6's own roster, auto-exclude (Type 5)
• "{Other Label} / EMPIRE" → surface as candidate with EMPIRE distribution flagged for context — your team decides
• "EMPIRE" standalone → pure EMPIRE indie distribution, treat as unsigned candidate

Kickoff input we'll need from you: any AB6-ownership-stake artists distributed by other labels that we should add to the structural denylist.

Two founders, no agency middleware. The person who scopes it builds it. Faster decisions, no telephone game.

Communication cadence

• Weekly written update every Monday — what shipped, what's blocked, what's next
• Bi-weekly Loom demo — 3-5 min screencast when there's visible progress
• Milestone meetings at kickoff, mid-point, pre-handoff, and handoff
• Channels: we come to your existing Slack / Monday / email setup — we don't make you adopt a new tool

All three share the same backend pipeline. The difference is interface and signal sources. Pricing in USD. Build delivered through Upwork, recurring infrastructure invoiced direct to AB6 cards by the service providers (no markup).

Recurring monthly costs (pass-through to AB6)

Billed by service providers directly to AB6. Node does not mark them up. Strategic tier adds ~$0-5/mo (Spotify playlist API and roster monitoring run on existing scrapers; no material increase).

Post-launch retainer — bug fixes only

This is a fixed-fee build. Once delivered, the system runs on its own. The retainer exists for one purpose: keeping it running when the unofficial endpoints we depend on change something internal.

What the retainer does NOT cover: new features, new platforms (YouTube Shorts, SoundCloud, Audiomack), business-logic changes, migrations or upgrades. These are quoted separately as Phase 2 work.

Pricing summary

Reply with your tier and post-launch preference, and we kick off this week. Kickoff call within 48 hours of acceptance — that's where you give us the scene reference list, confirm Monday board column preferences, lock the genre allowlist, and identify the Slack channel for notifications.

Kickoff readiness gate

Account provisioning needs to complete by end of week 1 or the timeline shifts proportionally. We come prepared with research-driven defaults so kickoff is decisions-only, not greenfield discovery — typical kickoff call is 60-90 minutes.

What we'll need from you at kickoff

• Scene reference list — ~30-50 rap/phonk/cloud-rap artist accounts that represent AB6's A&R taste. We come with a research-driven starter list; you sharpen.
• Initial filter thresholds — start permissive (per your “I want more” signal on the call), tune in week 4
• Slack channel routing — which channel, which team members get high-priority pings
• Monday board column preferences — we design a default board, you override any columns you want different
• Account provisioning — Apify, Vercel, Supabase, Railway, Anthropic, Perplexity. We send a provisioning checklist with direct signup links
• AB6-ownership-stake artist list (Strategic tier only) — any artists distributed by other labels that AB6 has ownership stake in, so we add them to the structural denylist