Real-time price intelligence,
sourced from the ground.

Query individual price records with rich filters. Supports pagination, sorting, and CSV export. Use this when you want raw data — for storage, custom analysis, or feeding your own pipeline.

Example: Hungarian dairy, last 30 days, sorted newest first

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/prices?country=HU&limit=5"

Sample response:

{
  "meta": {
    "total": 10265,
    "returned": 5,
    "limit": 5,
    "offset": 0,
    "tier": "starter",
    "tier_max_limit": 100,
    "cost_credits": 1,
    "queries_remaining": 9999,
    "response_ms": 142
  },
  "data": [
    {
      "id": 7633,
      "product_name": "MILK 1L",
      "product_brand": "Auchan",
      "category": "grocery",
      "subcategory": "dairy",
      "price": 349,
      "currency": "HUF",
      "price_usd": 0.96,
      "country": "HU",
      "city": "Budapest",
      "lat": 47.526462,
      "lng": 19.039627,
      "store_name": "Auchan Magyarország Kft.",
      "store_type": "supermarket",
      "store_chain": "Auchan",
      "confidence": 0.92,
      "source": "receipt",
      "captured_at": "2026-04-21T18:52:00+00:00"
    }
  ]
}

See Query Parameters for the complete filter list and Response Schema for every field in a record.

Aggregate statistics over the price database — median, average, percentiles, standard deviation, and a 7-day trend signal. Returns a single object (or grouped breakdown when group_by is set). Cheaper than fetching all records and aggregating client-side.

Parameters

All /prices filters work here, plus:period (7d, 30d, 90d, all), group_by (none, country, city, store_chain, subcategory, category — Growth+ for non-none), include_outliers (default false; set to true for raw stats including extreme values).

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/stats?period=30d"

Sample response — global aggregate over the last 30 days:

{
  "meta": {
    "credits_used": 2,
    "remaining": 9998,
    "period": "30d",
    "group_by": "none",
    "filters_applied": {},
    "response_ms": 988
  },
  "data": {
    "count": 8677,
    "avg_usd": 9.632,
    "min_usd": 0.05,
    "max_usd": 2150,
    "median_usd": 3.03,
    "p25_usd": 2.07,
    "p75_usd": 4.65,
    "stddev_usd": 103.4002,
    "latest_captured_at": "2026-04-21T18:52:00+00:00",
    "trend_7d_vs_prior_pct": -5.61,
    "countries": 14,
    "cities": 33,
    "store_chains": 10
  }
}

The trend_7d_vs_prior_pct field compares the median of the last 7 days against the prior 23 days. Useful as an early inflation signal.

Detect significant price changes by comparing a recent window against a baseline. Aggregates at the subcategory + country [+ city] level — surfaces regional inflation signals, supply disruptions, and competitive pricing shifts.

Parameters

recent_days (1-90, default 7), baseline_days (7-365, default 30, must be > recent_days), threshold_pct (1-100, default 10), direction (up, down, both), min_sample (default 5), group_by_city (default true), include_outliers (default false). Plus the standard filters: country, category, subcategory, city, store_chain.

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/anomalies?threshold_pct=10&recent_days=7"

Sample response — a real US anomaly detected in our database:

{
  "meta": {
    "credits_used": 5,
    "remaining": 9999995,
    "recent_window": {
      "days": 7,
      "start": "2026-04-16T10:20:04Z",
      "end": "2026-04-23T10:20:04Z"
    },
    "baseline_window": {
      "days": 23,
      "start": "2026-03-24T10:20:04Z",
      "end": "2026-04-16T10:20:04Z"
    },
    "threshold_pct": 10,
    "direction": "both",
    "results_count": 1,
    "tier_max_results": 500,
    "response_ms": 1147
  },
  "data": [
    {
      "category": "grocery",
      "subcategory": "beverages",
      "country": "US",
      "city": "Town of Smithtown",
      "recent_median_usd": 7.37,
      "baseline_median_usd": 5.69,
      "recent_avg_usd": 7.37,
      "baseline_avg_usd": 5.69,
      "change_pct": 29.53,
      "direction": "up",
      "recent_samples": 40,
      "baseline_samples": 20,
      "first_detected": "2026-04-21T00:00:00Z",
      "last_observed": "2026-04-21T00:00:00Z"
    }
  ]
}

Results are sorted by absolute change_pct descending. The recent_samples and baseline_samples fields let you assess statistical significance.

Time-bucketed aggregates ready for direct charting. Each bucket contains median, average, percentiles, and sample count. Choose daily, weekly, or monthly granularity.

Tier caps: Starter 90 buckets · Growth 365 · Pro 730 · Enterprise 1,825. Requesting beyond your cap automatically trims the period — no error, just a cap_applied: true flag in the response.

Parameters

period_days (1-1825), granularity (daily/weekly/monthly), min_sample (default 1, exclude buckets with fewer samples), include_outliers (default false; per-bucket IQR filter applied). Plus the standard filters: country, category, subcategory, city, store_chain.

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/timeseries?country=HU&subcategory=dairy&period_days=180&granularity=weekly"

Sample response — Hungarian dairy, weekly for 180 days:

{
  "meta": {
    "credits_used": 3,
    "remaining": 9999992,
    "granularity": "weekly",
    "period_days": 180,
    "period_start": "2025-10-25T10:38:23Z",
    "period_end": "2026-04-23T10:38:23Z",
    "filters_applied": {
      "subcategory": "dairy",
      "country": "HU"
    },
    "min_sample": 1,
    "buckets_returned": 2,
    "tier_max_buckets": 1825,
    "response_ms": 790
  },
  "data": [
    {
      "bucket_start": "2026-04-13T00:00:00Z",
      "samples": 12,
      "median_usd": 2.63,
      "avg_usd": 2.2825,
      "min_usd": 1.16,
      "max_usd": 3.48,
      "p25_usd": 1.425,
      "p75_usd": 2.9875,
      "stddev_usd": 0.8719
    },
    {
      "bucket_start": "2026-04-20T00:00:00Z",
      "samples": 17,
      "median_usd": 2.66,
      "avg_usd": 2.6141,
      "min_usd": 1.27,
      "max_usd": 4.93,
      "p25_usd": 1.38,
      "p75_usd": 3.42,
      "stddev_usd": 1.2281
    }
  ]
}

Buckets without samples are omitted (not zero-filled). Use min_sample=N to require minimum statistical significance per bucket.

Bulk download for analysis pipelines, hedge fund backtests, and academic research. Supports csv, ndjson (newline-delimited JSON, streamable), and json. Streams chunked — millions of rows without buffering memory.

Tier limits (max rows per call): Starter 10K · Growth 100K · Pro 1M · Enterprise 5M. JSON format capped at 100K rows; use ndjson or csv for larger exports.

Parameters

format (csv/ndjson/json), country, city, category, subcategory, data_source (collector/wfp/scrape/openfoodfacts), since, until, min_price_usd, max_price_usd, limit.

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/export?country=KE&format=csv&since=2025-01-01" \
  > kenya_2025.csv

Cost: 1 credit per 1,000 rows returned (rounded up, minimum 1). A 75K-row export = 75 credits.

Liksy Food Inflation Index — a monthly, country-specific basket cost normalized to a January 2020 baseline of 100. Methodology mirrors the WFP Minimum Food Basket framework with country-specific weight overrides, IQR outlier filtering, and unit normalization. Read the full methodology at liksy.app/methodology.

Coverage: 12 countries with WFP institutional data — KE, NG, ET, EG, YE, SY, LB, UA, JO, IQ, GH, TR. Each returns up to 130+ monthly index points. New countries added as ingestion expands.

Parameters

country (required, ISO2), months (history depth, default 24, max 360).

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/inflation?country=KE&months=24"

Sample response — Kenya, 24 months:

{
  "meta": {
    "country": "KE",
    "as_of": "2026-03-01",
    "tier": "starter",
    "cost_credits": 3,
    "response_ms": 943
  },
  "latest": {
    "index_value": 131.72,
    "basket_cost_usd": 1.773,
    "yoy_pct": 0.62,
    "mom_pct": 7.86
  },
  "methodology": {
    "basket_name": "standard",
    "baseline_date": "2020-01-01",
    "baseline_value": 100,
    "docs": "https://liksy.app/methodology"
  },
  "basket_components": {
    "effective_weight": 0.5,
    "subcategories_used": 2,
    "breakdown": [
      { "subcategory": "grains", "weight": 0.4, "current_median_usd_per_unit": 0.75, "yoy_pct": 0 },
      { "subcategory": "meat", "weight": 0.1, "current_median_usd_per_unit": 5.865, "yoy_pct": 8.21 }
    ]
  },
  "history": [
    { "month": "2024-04-01", "index": 123.36 },
    { "month": "2026-03-01", "index": 131.72 }
  ]
}

Index returns null when basket coverage falls below 40% (insufficient subcategory data). Always check basket_components.effective_weight for the subcategory share actually measured.

Composite 0–100 food-risk score with S&P-style letter grade (AAA→D). Designed for trade credit insurance, sovereign macro analysis, and EM hedge fund overlays. Combines inflation volatility (40%), absolute inflation (30%), anomaly density (20%), and data freshness (10%) over a 24-month rolling window.

Use compare to benchmark: add up to 9 additional countries with compare=NG,ET,YE. Returns the same risk breakdown for every country in a single call. 1 extra credit per compared country.

Parameters

country (required, ISO2), compare (comma-separated ISO2 list, max 9).

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/risk/country?country=KE&compare=NG,ET,YE,LB"

Sample response — Kenya benchmarked against 4 EM/crisis countries:

{
  "meta": {
    "country": "KE",
    "tier": "starter",
    "cost_credits": 9,
    "response_ms": 2858
  },
  "primary": {
    "country": "KE",
    "risk_score": 30.66,
    "letter_grade": "BBB",
    "components": {
      "volatility": { "risk_score_0_100": 27.4, "weight": 0.4 },
      "absolute_inflation": { "risk_score_0_100": 52.57, "weight": 0.3 },
      "anomaly_density": { "risk_score_0_100": 12, "weight": 0.2 },
      "data_freshness": { "risk_score_0_100": 15.33, "weight": 0.1 }
    },
    "data_quality": {
      "inflation_months_used": 11,
      "avg_index": 126.29,
      "stddev_index": 6.92,
      "latest_data_date": "2026-03-01"
    }
  },
  "methodology": {
    "formula": "0.40*volatility + 0.30*inflation + 0.20*anomalies + 0.10*freshness",
    "grades": { "AAA": "<10", "AA": "10-19", "A": "20-29", "BBB": "30-44", "BB": "45-59", "B": "60-74", "CCC": "75-89", "D": ">=90" },
    "docs": "https://liksy.app/methodology"
  },
  "comparison": [
    { "country": "NG", "risk_score": 38.01, "letter_grade": "BBB" },
    { "country": "ET", "risk_score": 14.67, "letter_grade": "AA" },
    { "country": "YE", "risk_score": 70.51, "letter_grade": "B" },
    { "country": "LB", "risk_score": 43.51, "letter_grade": "BBB" }
  ]
}

Countries with insufficient data (<6 months in 24-month window) return letter_grade: "NR" (Not Rated). Full methodology at liksy.app/methodology.

Pearson correlation coefficient between two countries' monthly median prices for a given subcategory. Reveals price passthrough (supply chain coupling), regional cohesion, and crisis dissociation. Two modes: pair (A ↔ B) and matrix (A vs many).

Real example: UA ↔ NG grains correlation = −0.81 over 24 months. When Ukraine grain exports tighten, Nigerian markets compensate through alternative sourcing — a negative correlation signals shock absorption, not transmission. Hedge fund alpha signal.

Parameters

Pair mode: country_a, country_b, subcategory, months (default 24, max 60).
Matrix mode: country, matrix (comma-separated ISO2 list, max 9), subcategory, months.

Example

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/correlation?country=UA&matrix=NG,EG,KE,YE,LB,JO,IQ,SY&subcategory=grains&months=24"

Sample response — Ukraine grains vs 8 EM countries, sorted by coupling strength:

{
  "meta": {
    "mode": "matrix",
    "anchor_country": "UA",
    "cost_credits": 12,
    "response_ms": 4850
  },
  "correlations": [
    {
      "country_a": "UA",
      "country_b": "NG",
      "subcategory": "grains",
      "pearson_r": -0.8069,
      "overlapping_months": 24,
      "interpretation": {
        "label": "very strong negative",
        "variance_explained_pct": 65,
        "note": "tight coupling — consider as single exposure"
      }
    },
    {
      "country_a": "UA",
      "country_b": "IQ",
      "pearson_r": -0.4696,
      "overlapping_months": 23,
      "interpretation": { "label": "moderate negative", "variance_explained_pct": 22 }
    }
  ],
  "methodology": {
    "method": "Pearson correlation coefficient",
    "aggregation": "monthly median USD price per unit (kg/L), IQR-trimmed",
    "sort_order": "absolute correlation strength, descending"
  }
}

Each month requires at least 3 IQR-trimmed samples per country to be included in the correlation. Minimum 6 overlapping months required for meaningful interpretation — fewer returns insufficient_data.

Instant answer to "what data do you have?" — designed for pre-sales demos, self-service prospect evaluation, and internal monitoring. Three modes:

• Global (no params): platform snapshot — total records, top countries, category breakdown, 30/7-day momentum.
• Country drill (?country=XX): subcategory + city + data_source breakdown with freshness status.
• Filtered (any combination): record count + status + sample rows for your exact query.

Coverage status labels: strong (5K+ records, active), partial (100-5K), sparse (<100), stale (no updates in 90 days), no_coverage (none). Honest signaling — we flag gaps explicitly.

Parameters

All optional: country, city, category, subcategory, product. No params = global snapshot (served from hourly cache, <200ms).

Example

# Global snapshot — no auth burden, 1 credit, cache served
curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/coverage"

# Country drill-down
curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/coverage?country=TR"

# Filtered query — "do you have grains in Nigeria?"
curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/coverage?country=NG&subcategory=grains"

Sample response — Türkiye country drill (stale status demonstration):

{
  "meta": {
    "mode": "country",
    "tier": "starter",
    "cost_credits": 1,
    "response_ms": 1240
  },
  "coverage": {
    "country": "TR",
    "status": "stale",
    "total_records": 11392,
    "total_cities": 4,
    "earliest_capture": "2013-05-15T00:00:00Z",
    "latest_capture": "2022-03-15T00:00:00Z",
    "median_age_days": "2535.9",
    "records_last_30d": 0
  },
  "by_subcategory": [
    { "subcategory": "vegetables", "records": 3092, "median_usd": 0.80 },
    { "subcategory": "condiments", "records": 1840, "median_usd": 5.70 },
    { "subcategory": "grains", "records": 1561, "median_usd": 0.88 }
  ],
  "by_city": [
    { "city": "Istanbul", "records": 2566, "latest_capture": "2022-03-15T00:00:00Z" },
    { "city": "Ankara", "records": 2599 },
    { "city": "Izmir", "records": 2590 }
  ],
  "status_legend": {
    "strong": "5,000+ records, actively updated",
    "stale": "no updates in last 90 days",
    "no_coverage": "no data for this country"
  }
}

Cheapest endpoint at 1 credit — designed to be queried freely during evaluation. Global mode served from hourly materialized cache for <200ms latency.

Supply Chain Stress Index. Detects when the same subcategory shows statistically significant price moves across multiple countries simultaneously — a leading indicator of upstream supply shocks, trade disruption, or regional commodity stress.

Real example: Current run detects oils surge in JO +16.7%, LB +13.0%, KE -13.5% — Middle East / East Africa price decoupling signal. Hedge fund pair trade setup, Lloyd's cargo risk repricing trigger.

Methodology

For every country × subcategory pair we compute the z-score of the recent median vs a preceding baseline distribution. When |z| exceeds the threshold in N+ countries simultaneously, the subcategory is flagged as a cluster and ranked by severity (avg |z| × countries affected).

Parameters

All optional: recent_days (14-365, default 90), baseline_days (60-1825, default 540), min_z (0.5-5.0, default 1.0), min_countries (2-10, default 2). Defaults tuned for WFP monthly-cadence data.

Example

Severity scale: critical (avg |z| ≥ 3.0 AND ≥ 5 countries), high, elevated, moderate. Each cluster includes per-country z-scores and automated interpretation narrative.

/api/v1/index2 CREDITSNEW

External authoritative index series — currently FAO Consumer Price Indices (food + general CPI) for 151 countries, January 2000 to present. Liksy serves upstream values verbatim under CC-BY-4.0. Use this to cross-check Liksy-computed inflation against official reference data, or as a standalone time-series feed for research dashboards.

Parameters

Example — Turkey food CPI, last 36 months

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/index?country=TR&series=food_cpi"

Example response (truncated)

{
  "meta": {
    "tier": "starter",
    "cost_credits": 2,
    "total_points": 29,
    "series_returned": 1,
    "response_ms": 142
  },
  "country": "TR",
  "time_window": { "from": "2023-04-24", "to": "2026-04-24" },
  "series": [{
    "data_source": "faostat_cp",
    "series_code": "food_cpi",
    "series_label": "FAO Food Consumer Price Index",
    "unit": "index",
    "base_period": "2015=100",
    "cadence": "monthly",
    "points": [
      { "period": "2023-05", "value": 660.90, "flag": "X", "confidence": 0.95 },
      { "period": "2024-09", "value": 1181.23, "flag": "X", "confidence": 0.95 },
      { "period": "2025-09", "value": 1607.19, "flag": "X", "confidence": 0.95 }
    ],
    "latest": { "period": "2025-09", "value": 1607.19 },
    "yoy_pct": 36.06
  }],
  "methodology": {
    "note": "External authoritative index series — Liksy serves upstream values verbatim, no recomputation.",
    "sources": {
      "faostat_cp": "FAO Consumer Price Indices, monthly, 2015=100. License CC-BY-4.0."
    },
    "docs": "https://liksy.app/methodology"
  }
}

Discover what's available for a country

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/index?country=KE&list_series=true"

/api/v1/webhooksFREE MGMTNEW

Real-time event delivery via HMAC-signed POST callbacks. Register a URL once, then receive push notifications when Liksy detects price anomalies or supply-chain stress clusters that match your filters — no polling, no missed events, no stale data.

Why webhooks instead of polling /anomalies every minute? You don't pay credits for empty polls, you receive events within a minute of detection (vs your polling interval), and Liksy's HMAC-SHA256 signature proves the payload originated from us so you can trust it.

How it works

Liksy runs a trigger every 15 minutes that recomputes anomalies and stress clusters across the database. Matching events are queued for each webhook subscription whose filters apply. A separate delivery worker drains the queue every minute, signing each payload with your webhook's secret and POSTing to your URL. Failed deliveries retry with exponential backoff (1m, 5m, 30m); after 3 failures the delivery is marked dead and the event is visible in your audit log.

Register a webhook

curl -X POST "https://liksy.app/api/v1/webhooks" \
  -H "Authorization: Bearer lk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Trading desk anomaly feed",
    "url": "https://your-server.example.com/liksy-webhook",
    "event_types": ["anomaly_detected", "supply_chain_stress"],
    "filter_country": "TR",
    "filter_subcategory": "dairy",
    "min_severity": "elevated"
  }'

The response contains a one-time secret (32-byte hex) — store it securely. You'll need it to verify signatures on incoming deliveries. Liksy never returns this secret again.

Sample response

{
  "id": "9f319edd-53d8-4004-bde5-2587cf00c89d",
  "name": "Trading desk anomaly feed",
  "url": "https://your-server.example.com/liksy-webhook",
  "secret": "3d3ceeda599a04961ed5fcd37d9378907ad79c75f9b157f053b564b8edf9144f",
  "secret_warning": "This secret is shown ONCE. Store it now — it cannot be retrieved later.",
  "event_types": ["anomaly_detected", "supply_chain_stress"],
  "filter_country": "TR",
  "filter_subcategory": "dairy",
  "min_severity": "elevated",
  "created_at": "2026-04-25T14:35:42.496Z"
}

List your webhooks

curl -H "Authorization: Bearer lk_live_YOUR_KEY" \
  "https://liksy.app/api/v1/webhooks"

Returns all your subscriptions with delivery counters (total_deliveries, total_failures, last_triggered_at) but never the secret. Revoked subscriptions remain visible with is_active: false for audit.

Revoke a webhook

curl -X DELETE "https://liksy.app/api/v1/webhooks/{id}" \
  -H "Authorization: Bearer lk_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "reason": "rotated to new endpoint" }'

Soft delete — the subscription is marked inactive immediately, no new events are queued, but pending deliveries already in the queue will still be sent before the worker picks up the revocation. Idempotent: revoking an already-revoked webhook returns 200.

Filtering

Incoming delivery — request headers

Every delivery POST to your URL includes these headers:

Sample payload — anomaly_detected

{
  "event": "anomaly_detected",
  "severity": "high",
  "country": "TR",
  "category": "grocery",
  "subcategory": "dairy",
  "city": null,
  "change_pct": 31.8,
  "direction": "up",
  "recent_avg_usd": 1.45,
  "baseline_avg_usd": 1.10,
  "recent_median_usd": 1.42,
  "baseline_median_usd": 1.08,
  "recent_samples": 87,
  "baseline_samples": 412,
  "first_detected": "2026-04-22T08:14:00Z",
  "last_observed": "2026-04-25T11:58:54Z",
  "detected_at": "2026-04-25T14:48:00Z"
}

Sample payload — supply_chain_stress

{
  "event": "supply_chain_stress",
  "severity": "elevated",
  "subcategory": "oils",
  "countries_affected": 3,
  "country_list": ["JO", "LB", "KE"],
  "avg_price_move_pct": 14.4,
  "avg_z_score": 1.85,
  "max_z_score": 2.31,
  "direction": "up",
  "severity_score": 5.55,
  "detected_at": "2026-04-25T14:48:00Z"
}

Verifying the signature — Node.js

import crypto from "node:crypto";
import express from "express";

const app = express();

// IMPORTANT: capture the raw body BEFORE JSON parsing.
// HMAC verification depends on the exact byte sequence.
app.post("/liksy-webhook",
  express.raw({ type: "application/json" }),
  (req, res) => {
    const signature = req.header("X-Liksy-Signature") || "";
    const expected = "sha256=" + crypto
      .createHmac("sha256", process.env.LIKSY_WEBHOOK_SECRET)
      .update(req.body)
      .digest("hex");

    // Constant-time compare avoids timing attacks
    const ok = signature.length === expected.length &&
      crypto.timingSafeEqual(Buffer.from(signature), Buffer.from(expected));

    if (!ok) return res.status(401).send("invalid signature");

    const event = JSON.parse(req.body.toString());
    const deliveryId = req.header("X-Liksy-Delivery-Id");

    // Idempotency: skip if you've seen this delivery_id before

    console.log("Liksy event:", event.event, event.severity);
    res.status(200).send("ok");
  }
);

Verifying the signature — Python (Flask)

import hmac, hashlib, os
from flask import Flask, request, abort

app = Flask(__name__)
SECRET = os.environ["LIKSY_WEBHOOK_SECRET"].encode()

@app.post("/liksy-webhook")
def liksy_webhook():
    raw_body = request.get_data()  # exact bytes, BEFORE Flask parses JSON
    signature = request.headers.get("X-Liksy-Signature", "")
    expected = "sha256=" + hmac.new(SECRET, raw_body, hashlib.sha256).hexdigest()

    if not hmac.compare_digest(signature, expected):
        abort(401, "invalid signature")

    event = request.get_json()
    delivery_id = request.headers.get("X-Liksy-Delivery-Id")

    print(f"Liksy event: {event['event']} {event['severity']}")
    return ("ok", 200)

Retry policy

Liksy considers a delivery successful only on a 2xx response within 8 seconds. Any other status (including 5xx, redirects, timeouts) triggers a retry. Backoff schedule is +1 min, then +5 min, then +30 min. After 3 failed attempts the delivery is marked dead and surfaced in your operator console. For idempotency, dedupe on X-Liksy-Delivery-Id — the same delivery may arrive twice if your server times out after processing but before responding.

Idempotency & deduplication

The same anomaly cluster (e.g. "TR dairy high") can recur across multiple 15-minute trigger runs while the underlying condition persists. Liksy fingerprints each event by (country, subcategory, severity, direction, day) and de-duplicates within a 24-hour window — you receive at most one notification per webhook per event-day. The fingerprint key is internal; you should still dedupe on X-Liksy-Delivery-Id for retry safety.

Webhook management endpoints (POST/GET/DELETE) do not consume credits and do not count against your monthly quota. Event deliveries are also free — you only pay for queries you issue against the data API.