# Changelog of TikTok Product Intelligence (`beastmasteral/tiktok-product-intelligence`) Actor

- **URL**: https://apify.com/beastmasteral/tiktok-product-intelligence/changelog.md
- **Full Actor documentation**: https://apify.com/beastmasteral/tiktok-product-intelligence.md

## Changelog

### Store publication limited-permissions patch - 2026-09-29

- Removed in-Actor Apify Schedule mutation from the public Store path.
- Monitoring now keeps durable same-scope alert state; recurring execution is configured externally in Apify Schedules.
- Migrated default history and monitoring Key-Value Store names to new `*-limited-v1` stores for clean Limited-permissions ownership.
- Reduced Store default/prefill workload to 5 organic posts, 3 comments per post, and 5 Shop products for fast daily Store health tests.
- Preserved M14 evidence, matching, supplier, landed-cost, and PPE semantics.

### Store publication monetization patch - 2026-09-29

- Added guarded PPE charging for `core-analysis`, `ads-analysis`, and `supplier-analysis`.
- Charging activates only when Apify reports the Actor is running under Pay-per-event pricing.
- Charges are emitted only after Store output has been persisted.
- Optional Ads/Supplier events are not charged when their stage is disabled or skipped.
- No M14 evidence, matching, supplier, or landed-cost semantics changed.

### 1.2.0 - M14 Supplier Verification Evidence & Landed-Cost Scenarios

- Added public supplier due-diligence evidence for product-linked supplier offers.
- Preserved public trust/profile badges, reviews, response rate and certifications without claiming independent verification.
- Added explicit `NOT_VERIFIED_BY_TPI` conclusion to supplier verification evidence.
- Added opt-in landed-cost scenarios using observed supplier unit cost plus explicit user-declared per-unit import costs.
- Added destination country and common-currency contract; no FX conversion.
- Added fail-closed input validation with no hidden numeric defaults for freight/duty/tax/other import costs.
- Added transparent landed-cost and retail-minus-landed-cost formulas.
- Kept profitability `UNKNOWN` and avoided customs/tax/legal determinations.
- Extended Store schemas, sample output, readiness audit, fixture tests and package verifier.
- Preserved M1-M13 semantics and modularity.

### 1.1.0 - M13 Supplier Sourcing & Cost Evidence

- Added opt-in supplier sourcing with configurable upstream Actor.
- Added public Alibaba offer normalization: price ranges, quantity tiers, MOQ, supplier/trust evidence, certifications and provenance.
- Added conservative supplier-to-product linking; generic/brand-only/ambiguous offers remain unresolved.
- Added quantity-aware observed supplier unit-cost selection.
- Added same-currency retail-minus-supplier-unit-cost `grossSpreadProxy`.
- Added partial landed-cost proxy only for explicit observed per-unit shipping.
- Kept profitability `UNKNOWN` without a complete landed/selling cost stack.
- Added independent fail-closed supplier row budget.

### 1.0.0 - M12 Store-Ready MVP

- Published the accepted M1-M11 pipeline as a coherent Store-ready release surface.
- Added Store-facing README, sectioned input UI, output schemas, dataset overview, compact `RUN_SUMMARY.json`, release audit and publication checklist.
- Preserved evidence-blocking statuses and uncertainty boundaries.