NSQ Watch - CDSCO NSQ drug alerts and stock batch matcher

NSQ Watch is currently paused by its publisher while permission for the source data is being clarified. Every run ends immediately with one result row whose status is paused. No data is fetched from CDSCO or any other website, your input is not read, nothing is charged, and the run succeeds. The description below explains what the Actor does when it is active. Our privacy policy and EULA remain in force.

NSQ Watch turns the "Not of Standard Quality" (NSQ), state NSQ and spurious drug lists of India's drug regulator CDSCO into clean rows, and checks your pharmacy or distributor stock list against the flagged batches every day. Set it up once as a scheduled Task: when CDSCO publishes new rows, you get a quarantine list of the stock lines whose batch appears in them.

It is built for pharmacy chains and e-pharmacies, hospital pharmacy and procurement teams, distributors and stockists, pharma quality and regulatory teams, health-tech apps, and researchers or journalists who want the lists as data.

NSQ Watch reports what the official CDSCO publication lists. It is a screening signal, not medical or legal advice. A flagged batch means a tested sample failed; always verify with the official publication and your supplier before you act.

Independent tool, not affiliated with or endorsed by CDSCO. The CDSCO Copyright Policy (https://cdsco.gov.in/opencms/opencms//en/Copyright-Policy/) states: "The contents of this website may not be reproduced partially or fully, without due permission from Central Drugs Standard Control Organization, Govt. of India. If referred to as a part of another publication, the source must be appropriately acknowledged. The contents of this website can not be used in any misleading or objectionable context." The publisher has not asked CDSCO for permission. NSQ Watch fetches the public data inside your own Apify account and links every row to its official source; if you republish the rows or build a product on them, you are responsible for asking CDSCO for permission, acknowledging CDSCO as the source and not using the data in a misleading context.

Why use it

• Machine-readable NSQ lists. One row per flagged batch: product (with brand when printed), batch number, manufacturing and expiry month, manufacturer and address, the test result (reason), the testing laboratory or state, alert month, and a link to the official source.
• Stock matching with typo tolerance. Your stock lines are compared with the flagged batches by batch number, product name and manufacturer. Batch numbers are compared with spaces and punctuation ignored, and common letter/digit confusions (O and 0, I and 1, S and 5, B and 8, Z and 2) are tolerated. Every match shows its score and the reasons.
• Quarantine CSV. Matches are written with your own columns (sku, location, quantity) first, so the file can go straight into your system.
• Only new things are billed and delivered. A row or match you already received is never delivered or charged again. A run without new alerts costs only the small daily check.
• No server, no proxy bill for you. Plain HTTP and PDF reading; your stock list stays in your own Apify account.

Sources that are monitored

Source	What it provides
CDSCO NSQ portal, https://cdscoonline.gov.in/CDSCO/viewPublicNSQDrug (the public data the page displays)	NSQ drugs from CDSCO laboratories and state laboratories, and spurious drugs, by reporting month
CDSCO Alerts page, https://cdsco.gov.in/opencms/opencms/en/Alerts/	PDF lists. Used for months the portal does not cover, and when the portal cannot be reached

Nothing else is monitored: no state drugs-control websites, no medical device alert parsing, no other countries. NSQ Watch is not affiliated with or endorsed by CDSCO. Requests follow robots.txt and are limited to one per second per host; no login, no CAPTCHA and no form is ever used.

How to use it

1. Click Try for free and keep the prefilled input. The first run is short and shows the format of the results (a sample stock line will normally not match anything).
2. Put your stock in the input: a few lines in Stock items (quick entry), or a link in Stock list CSV (URL) to a CSV exported from your billing software (Marg, GoFrugal or similar). The CSV needs a header row; common header names such as Item Name, Product Name, Batch No, Mfg, Company, Expiry and Qty are recognised, others can be mapped in CSV column names.
3. Choose Check alerts from the last (default 6 months) and press Save as a new task.
4. Open the task, add a Schedule: for example 0 11 * * * with timezone Asia/Kolkata (every day 11:00 IST), and enable it. Schedules start disabled, and an Actor must have been run once before it can be scheduled.
5. In the task's Integrations tab add Slack, email (Gmail) or a webhook on Run succeeded (see Notifications), and optionally one on Run failed.

Input

Field	What it does
mode	match (default) checks your stock; rows delivers every new alert row as data
inventory	Quick entry, up to 500 lines: PRODUCT | BATCH: x | Mfr: y (batch and Mfr optional; Exp:, SKU:, Loc:, Qty: also accepted)
inventoryCsvUrl	https link to a CSV, up to 50,000 rows and 20 MB. Google Drive share links and Dropbox links are converted to direct downloads
columnMap	Your CSV header for each field: product, batch, manufacturer, expiry, sku, location, quantity
lookbackMonths	0 to 24, default 6. Used on the first run and whenever stock lines change
matchLevel	batch (default) or product (also reports same product and maker with another batch, as warnings)
minScore	60 to 100, default 80
alertTypes	NSQ_CENTRAL, NSQ_STATE, SPURIOUS, RECALL_ALERT (all by default)
deliverAlertRows	Also deliver every new alert row (charged per row). Always on in rows mode
maxResultsPerRun	1 to 100,000, default 500
webhookUrl, slackWebhookUrl	Optional https endpoints notified at the end of the run
useSharedCache, includeLegacyPdfs, stateKey	Advanced. useSharedCache is off by default and the Actor reads the official CDSCO sources directly

Sample input:

{
  "mode": "match",
  "inventory": ["Paracetamol Tablets IP 500 mg | BATCH: XYZ123 | Mfr: Example Pharma"],
  "lookbackMonths": 6,
  "matchLevel": "batch",
  "minScore": 80
}

Your stock list is read while the run is in progress, kept in memory, never written to the log, and written only to your own results (the matches, REPORT, QUARANTINE.csv and OUTPUT). Use a link that cannot be guessed, or a public record of an Apify key-value store.

Output

The default dataset contains these record types (the recordType field):

• inventory-match: one per match, with your stock line, the flagged alert row, matchLevel (batch or product), score and reasons.
• alert-row: one per new alert row (rows mode, or Also deliver new alert rows).
• alert-document: one per new CDSCO list found in the run and how it was read (not charged).
• run-summary: one per run, always the last item (not charged): status, counts, charged events, sources with error codes, warnings.

Dataset views: Stock matches, Alert rows, Alert documents, Run summary. A view shows the columns of its own record type and empty cells for the others; filter on recordType through the API when you need only one type.

Key-value store records: OUTPUT (JSON summary with the first 200 matches), REPORT (a standalone HTML page with a red banner when a batch-level match exists, the matches, product-level warnings and the new documents), QUARANTINE.csv.

Illustrative sample of a match (invented names):

{
  "recordType": "inventory-match",
  "matchLevel": "batch",
  "score": 100,
  "reasons": ["batch-exact", "product-strong", "mfr-strong"],
  "stock": {"product": "Veloxapril Tablets IP 10 mg", "batch": "ZN26A101", "manufacturer": "Zentora Lifesciences Pvt. Ltd.", "location": "Main store", "quantity": "240"},
  "alert": {
    "alertType": "NSQ_CENTRAL", "alertMonth": "2026-09",
    "product": "Veloxapril Tablets IP 10 mg (Veloxa-10 Tablets)", "batch": "ZN26A101",
    "mfgDate": "2026-01", "expDate": "2027-12", "manufacturer": "Zentora Lifesciences Pvt. Ltd.",
    "reason": "The sample does not conform to IP in respect of Assay", "reportedBy": "CDL, Kolkata",
    "pdfUrl": "https://cdscoonline.gov.in/CDSCO/viewPublicNSQDrug"
  },
  "detectedAt": "2026-10-24T05:40:12Z"
}

How matching works

A batch-level match needs a batch hit and a total score of at least minScore.

Signal	Points
Batch identical (spaces, case and punctuation ignored)	60
Batch identical after folding O/0, I/L/1, S/5, B/8, Z/2	55
Batch one edit apart (both at least 6 characters)	40
Product name strong (similarity 90 or more) / partial (75 or more)	25 / 15
Strength differs (for example 10 mg and 20 mg)	-15
Manufacturer strong (90 or more) / partial (75 or more) / different (below 50)	+15 / +8 / -10
Expiry month equal	+5

Examples: identical batch with a strong product name scores 85, with the same manufacturer 100. An identical batch on a different product scores 60 or less (50 when the makers also differ) and is not reported at the default minimum score, because batch numbers repeat across manufacturers. A one-edit batch is dropped when both sides have an expiry month and the months differ (a neighbouring batch, not the flagged one). A minimum score of 86 or more accepts only identical or letter/digit-confusion batch matches.

Product names are compared as sets of words without dosage-form words and strengths, so word order, & against and, and the brand in brackets do not matter. A stock name that is only one ingredient of a printed combination product is capped at "partial" so that single-ingredient stock is not matched to a different combination product from the same batch number alone.

With matchLevel = product, items without a batch, and rows whose batch does not match, can still be reported as warnings when the product and the manufacturer agree (score 35 to 79). They are always labelled product.

Pricing

NSQ Watch is pay-per-event: you pay only for what a run does, with no subscription. The table shows the prices for the Apify Starter plan; Scale/Business plans get 10% / 20% off.

Event	When it is charged	Price (Starter)
Actor start	Once per run, by Apify	Apify's standard start event
watch-check (Daily alert check)	Once per run, after your input is accepted	$0.03
inventory-item-checked (Stock item checked)	Per stock line checked, only in runs that have new alert rows or changed stock lines	$0.0005
inventory-match (Stock match found)	Per new match delivered	$0.05
alert-row (Alert row delivered)	Per new alert row delivered	$0.005

Monthly cost = runs x check price + (runs with new alert rows) x stock lines x item price + matches x match price + delivered rows x row price. Runs per month: daily 30, weekdays 22, weekly 4.33.

Example	Stock lines	Runs with new rows	Matches	Rows delivered	Monthly (Starter)
One pharmacy, daily	1,500	4	1	0	30 x $0.03 + 6,000 x $0.0005 + 1 x $0.05 = $3.95
Chain, daily	20,000	4	6	0	$0.90 + 80,000 x $0.0005 + 6 x $0.05 = $41.20
Data user, rows only	0	-	0	400	$0.90 + 400 x $0.005 = $2.90

The first run additionally checks every stock line once against the lookback window (1,500 lines cost $0.75). How many runs have new rows depends on how CDSCO publishes: the old Alerts page released each monthly list on a single day, while the NSQ portal shows a "Pending States" button for the current month, so a month can receive rows over several days. For a large stock list a weekly schedule limits the number of checks.

Set "Maximum cost per run" to a value that covers the checks. If your limit cannot pay for all stock-line checks of a run, NSQ Watch checks nothing and says so in the summary (stockItemChecksNeeded) instead of checking half. If it runs out while delivering matches or rows, it stops exactly there, saves its progress and delivers the rest in the next run; nothing is delivered or charged twice.

Free plan (trial): stock up to 25 lines, at most 5 matches and 10 alert rows per run, and a 1-month lookback. The run still succeeds, and the summary names the cap that applied (freePlanCaps).

Notifications

Add integrations in the task's Integrations tab on the event Run succeeded.

• Slack message template:

  NSQ Watch run {{resource.status}}. Results: https://console.apify.com/storage/datasets/{{resource.defaultDatasetId}}
  Run: https://console.apify.com/actors/runs/{{resource.id}}
  

• Email (Gmail integration): attach the REPORT or QUARANTINE.csv record of the run's key-value store.
• Webhook: or set webhookUrl in the input; the OUTPUT JSON is posted when the run ends. slackWebhookUrl posts a short message with up to 10 batch-level matches.
• Add the same on Run failed so that you notice a failed run.

Limitations and disclaimers

• Lists are sample-specific: a flagged batch means a tested sample failed. The Actor reports what CDSCO publishes and does not interpret it. It is not medical or legal advice. Verify with the official publication and your supplier before action.
• Stock lines are compared by text. Different brand names for the same product, or a stock name that shares no word with the printed name, can be missed; wrong matches are possible when names are very short. Check the reasons and score of every match.
• Rows read from a PDF without table lines, or by OCR, are marked lowConfidence. OCR for scanned pages is an optional fallback; a scanned page that cannot be read is reported as a warning, never skipped silently. Layout changes at the source appear as warnings in the run summary.
• Dates are month precision, as printed. State names are inferred from the laboratory and are empty for CDSCO central laboratories.
• The Actor depends on the availability and format of the CDSCO websites. If a source is down, the run still succeeds, the summary lists the source with an error code, and the missing part follows in a later run.
• Data source: CDSCO (Central Drugs Standard Control Organization). Every row links to the official source; please acknowledge the source when you republish the data. Copyright: the CDSCO Copyright Policy says its website contents "may not be reproduced partially or fully, without due permission" (quoted in full at the top). This Actor is not affiliated with or endorsed by CDSCO, does not republish CDSCO data itself (the shared cache is switched off), and does not use CDSCO material in a misleading context. Permission for any republication is your responsibility.
• Disclaimer: NSQ Watch is an automated screening aid. It can miss batches or match wrongly. It is not medical, pharmaceutical or legal advice and does not replace checking the official CDSCO publication and your supplier.

FAQ

Which lists are covered? CDSCO NSQ lists (central and state laboratories) and spurious drug lists from the CDSCO NSQ portal, and the PDF lists on the CDSCO Alerts page for months the portal does not cover.

Do I need a stock list? No. Set the mode to rows and every new alert row is delivered as data.

Does it work with brand names? It compares your product name with the printed name and with a brand or generic name in brackets, ignoring word order and dosage-form words. A name with no word in common with the printed name is not matched.

Why do I see "product" matches? They appear with matchLevel = product and mean the same product and maker with another batch. They are warnings, never counted as batch-level matches.

Is a rerun charged again? No. A run with the same input and no new alerts charges only the daily check. Rows and matches already delivered are remembered in a store named leons-nsqwatch-state in your account; delete it to start from scratch.

What if I change my stock list? Only new or changed lines are checked against the lookback window, and only those are charged for it. Editing quantities does not make a line new.

Is my stock list stored by the publisher? No. It stays in your account and the run. The privacy policy is at https://raw.githubusercontent.com/leons0016/ADApps/main/privacy/P4-NSQWatch.txt

Why was nothing checked and the status is SPENDING_LIMIT? Your maximum cost per run was too low for the stock-line checks; raise it (see Pricing).

Can I use it through the API? Yes: run the Actor leons0016/cdsco-nsq-drug-alerts with the same JSON input and read the dataset.

Changelog and support

See the Changelog in the Information tab. Questions and bug reports: use the Issues tab of this Actor, or write to atulhritik@hotmail.com. Please do not include confidential stock data in an issue.

Third-party licences and attributions: see docs/THIRD_PARTY_NOTICES.txt in the Actor source.