PubChem Compounds Scraper

Resolve PubChem compounds in batches by CID, chemical name, SMILES, InChI, or molecular formula. The Actor calls the official anonymous PubChem PUG REST service and exports normalized compound identity, structure identifiers, molecular properties, bounded synonyms, input-resolution context, and source provenance. No PubChem API key, browser, account, or proxy is required.

Use it to turn recurring identifier lists into integration-ready chemistry records without building request pacing, retry handling, one-to-many formula resolution, normalization, or Apify dataset delivery yourself.

What does this Actor do?

For every submitted identifier, the Actor:

1. validates the declared identifier type;
2. resolves it to one or more PubChem CIDs;
3. bounds one-to-many matches with maxResultsPerInput;
4. retrieves normalized identity, structure, and molecular properties;
5. retrieves up to maxSynonyms synonyms;
6. preserves the original input and match rank;
7. emits explicit missing-result or lookup-failure status rows; and
8. links every resolved record to its PubChem page and official API endpoint.

The result is one default Apify dataset suitable for JSON, CSV, Excel, API, webhook, and automation integrations.

Who is it for?

• Cheminformatics teams enriching structure registries or assay inputs.
• Pharmaceutical researchers normalizing names and structure identifiers to authoritative CIDs.
• Data engineers cleaning mixed identifier columns before warehouse ingestion.
• Laboratory operations teams adding formula, mass, structure, and synonym context to sample lists.
• Toxicology and materials researchers preparing reproducible compound datasets.
• Developers who want managed PubChem API pacing, retries, datasets, schedules, and webhooks.

This Actor resolves compound records. It does not scrape PubChem BioAssay data, patents, literature, vendor offers, or perform similarity, substructure, or biological-activity analysis.

Why use this PubChem compound lookup?

A single run can mix identifier namespaces while preserving exactly how every output was resolved. The output includes both machine identifiers and practical molecular properties. Formula searches are explicitly bounded because one formula can represent many compounds. Synonym requests degrade safely: if only synonyms fail, the useful property record is still returned. Transient network errors, rate limits, and temporary PubChem failures use bounded retries.

The implementation is HTTP-only, runs at 256 MB, and does not add browser or proxy overhead.

Supported identifiers

Type	Example	Resolution behavior
cid	2244	Uses the exact PubChem Compound ID
name	aspirin	Resolves a chemical, common, or registry name through PubChem
smiles	CC(=O)OC1=CC=CC=C1C(=O)O	Uses the official form-encoded SMILES route
inchi	InChI=1S/CH4/h1H4	Uses the official form-encoded InChI route
formula	C9H8O4	Returns bounded ranked matches from PubChem fast formula search
auto	2244 or caffeine	Detects numeric CIDs; other values are treated as names

Declare smiles, inchi, and formula explicitly. Automatic mode intentionally does not guess among these potentially ambiguous syntaxes.

Data you receive

Each dataset row includes:

• original inputIndex, inputValue, and resolved inputType;
• resolutionRank for one-to-many matches;
• resolution status for successful, missing, or failed lookups;
• PubChem cid, preferred title, and iupacName;
• molecularFormula, molecularWeight, exactMass, and monoisotopicMass;
• canonicalSmiles, isomericSmiles, inchi, and inchiKey;
• xlogp, tpsa, and PubChem complexity;
• formal electrical state and hydrogen-bond donor/acceptor counts;
• rotatable-bond, heavy-atom, and covalent-unit counts;
• a bounded synonyms array;
• sourceUrl, sourceApiUrl, and retrievedAt; and
• a bounded error message when resolution does not succeed.

Fields can be null when PubChem does not report a value for a compound.

How to run the Actor

1. Open the Actor in Apify Console.
2. Add one or more objects to Compounds to resolve.
3. Select the identifier type for each value.
4. Set the output, per-input match, and synonym limits.
5. Click Start.
6. Open the Dataset tab to inspect or export results.

Start small when testing broad molecular formulas, because a formula may map to many CIDs.

Input parameters

Field	Type	Default	Limit	Description
compounds	array	sample mixed batch	1–1,000	Objects containing value and an identifier type
maxItems	integer	100	1–5,000	Maximum resolved or unresolved dataset rows
maxResultsPerInput	integer	5	1–25	Maximum CIDs retained for each submitted input
maxSynonyms	integer	20	0–100	Maximum synonyms per resolved compound; zero skips synonym calls

Example input:

{
  "compounds": [
    { "value": "aspirin", "type": "name" },
    { "value": "2244", "type": "cid" },
    { "value": "CC(=O)OC1=CC=CC=C1C(=O)O", "type": "smiles" },
    { "value": "C8H10N4O2", "type": "formula" }
  ],
  "maxItems": 10,
  "maxResultsPerInput": 3,
  "maxSynonyms": 10
}

Output example

This abbreviated row comes from a real current run:

{
  "inputIndex": 0,
  "inputValue": "aspirin",
  "inputType": "name",
  "resolutionRank": 1,
  "status": "resolved",
  "cid": 2244,
  "title": "Aspirin",
  "iupacName": "2-acetyloxybenzoic acid",
  "molecularFormula": "C9H8O4",
  "molecularWeight": 180.16,
  "canonicalSmiles": "CC(=O)OC1=CC=CC=C1C(=O)O",
  "inchiKey": "BSYNRYMUTXBXSQ-UHFFFAOYSA-N",
  "xlogp": 1.2,
  "tpsa": 63.6,
  "synonyms": ["aspirin", "ACETYLSALICYLIC ACID", "50-78-2"],
  "sourceUrl": "https://pubchem.ncbi.nlm.nih.gov/compound/2244",
  "error": null
}

An unresolved input still produces a traceable row, but has null compound fields and is not charged as a resolved compound.

How much does it cost to resolve PubChem compounds?

Pricing uses pay per event:

• a $0.002 run-start fee covers fixed startup work; and
• each successfully resolved compound uses the tiered Compound resolved event.

At the BRONZE tier, the target price is $0.0016 per resolved compound. The exact active tier shown in Apify Console always controls billing. Unresolved status rows have no compound-resolution event fee.

BRONZE billing examples use this formula:

run total = $0.002 start + resolved compound count × $0.0016

For one, ten, one hundred, or one thousand useful compounds, multiply the same active per-compound rate by the resolved count and add the one-time active start fee. Missing or failed lookup rows do not add item events.

Higher Apify subscription tiers receive the lower event rates displayed on the Actor page.

Identifier resolution details

A name or structure can resolve differently as PubChem updates its records. The Actor preserves inputValue, inputType, and resolutionRank so downstream jobs can audit the mapping. CID lookup is the most deterministic mode when you already know the compound identifier.

Formula lookup is naturally one-to-many. Use maxResultsPerInput to select the number of ranked CIDs retained. The Actor does not claim that the first formula match is uniquely correct for your scientific context.

Reliability, pacing, and retries

The Actor paces official API requests below PubChem's published short-term request ceiling. Network failures, HTTP 429, and temporary 5xx responses are retried up to three times with backoff. Each request has a 30-second timeout. Permanent invalid inputs and stable not-found results are not retried blindly.

If a synonym endpoint fails after the property endpoint succeeds, the row remains resolved with an empty synonyms array. This avoids discarding useful identity and structure data because an optional enrichment failed.

Limits and failure behavior

• A run accepts at most 1,000 submitted identifiers.
• Output is capped at 5,000 rows.
• One input can produce at most 25 matches.
• Synonyms are capped at 100 per compound.
• maxItems stops the run even when more formula matches exist.
• PubChem data may be corrected, merged, deprecated, or expanded over time.
• Some compounds do not expose every requested property.
• This Actor does not validate laboratory purity, hazards, dosage, or clinical suitability.

A completely malformed run input fails with a non-zero status. An individual missing or failed identifier produces a status row so the rest of the batch can continue.

Batch enrichment workflow

A practical recurring pipeline is:

1. export identifiers from a LIMS, ELN, inventory, or warehouse;
2. map each source column to the correct identifier type;
3. run this Actor with an output bound appropriate to the batch;
4. join results using inputIndex or inputValue;
5. retain cid, inchiKey, and provenance as stable normalization fields;
6. review one-to-many formula matches using resolutionRank; and
7. schedule the same Task when periodic PubChem refreshes are required.

Apify schedules, webhooks, dataset exports, and integrations can automate the downstream steps.

API usage with cURL

Set APIFY_TOKEN, then run:

curl -X POST \
  "https://api.apify.com/v2/acts/automation-lab~pubchem-compound-batch-lookup/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"compounds":[{"value":"aspirin","type":"name"}],"maxItems":1,"maxResultsPerInput":1,"maxSynonyms":10}'

Use /run-sync-get-dataset-items instead of /runs when a caller should wait for the dataset response.

API usage with JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('automation-lab/pubchem-compound-batch-lookup').call({
  compounds: [
    { value: '2244', type: 'cid' },
    { value: 'C8H10N4O2', type: 'formula' },
  ],
  maxItems: 5,
  maxResultsPerInput: 2,
  maxSynonyms: 10,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

API usage with Python

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ['APIFY_TOKEN'])
run = client.actor('automation-lab/pubchem-compound-batch-lookup').call(run_input={
    'compounds': [
        {'value': 'caffeine', 'type': 'name'},
        {'value': 'InChI=1S/CH4/h1H4', 'type': 'inchi'},
    ],
    'maxItems': 2,
    'maxResultsPerInput': 1,
    'maxSynonyms': 5,
})
for item in client.dataset(run['defaultDatasetId']).iterate_items():
    print(item)

Use with MCP and AI agents

Add the Actor to Claude Code through Apify MCP:

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=automation-lab/pubchem-compound-batch-lookup"

Claude Desktop, Cursor, and VS Code can use this MCP server configuration:

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=automation-lab/pubchem-compound-batch-lookup"
    }
  }
}

Example prompts showing MCP usage include:

• “Resolve these CIDs and return formula, molecular weight, SMILES, and InChIKey.”
• “Normalize this list of chemical names to PubChem CIDs and keep source links.”
• “Find up to three PubChem compounds for each formula and compare exact masses.”

Always review generated scientific interpretations independently; the Actor returns source data, not scientific advice.

Legality and responsible use

PubChem is a public scientific resource operated by the U.S. National Library of Medicine. Use its records in accordance with PubChem policies and cite PubChem where appropriate. Keep sourceUrl, sourceApiUrl, and retrievedAt when reproducibility matters.

Compound records are informational. Do not use this Actor as the sole basis for clinical, safety, regulatory, or laboratory decisions. Validate critical identifiers and properties against authoritative domain procedures.

Troubleshooting

Why did a formula return several rows?

Molecular formulas are not unique identifiers. Reduce maxResultsPerInput, inspect resolutionRank, and use SMILES, InChI, or CID when you need structure-specific resolution.

Why does a row say the compound was not found?

Check the identifier namespace and spelling. Automatic mode only distinguishes numeric CIDs from names. Declare formulas and structure identifiers explicitly.

Why are synonyms empty on a resolved row?

The synonym request may have returned no values or failed after properties succeeded. Increase nothing blindly; rerun later or use the recorded CID to inspect PubChem directly.

Why did the whole run fail?

The Actor fails closed for an empty compounds list, more than 1,000 inputs, invalid bounds, blank values, or an unsupported identifier type. Inspect the run log for the indexed validation error.

FAQ

Does it require a PubChem key or login?

No. It uses official anonymous PUG REST routes.

Does it use a proxy or browser?

No. Direct structured HTTPS requests are sufficient for this workflow.

Are unresolved rows charged as compounds?

No. Only records with status: resolved emit the Compound resolved event. The one-time run-start fee still applies.

Can I submit duplicate identifiers?

Yes. The Actor preserves every input position. This is useful when joining output back to the original list.

Can it perform similarity or substructure search?

No. The supported workflow is exact identifier/formula resolution and record enrichment.

Can results change?

Yes. PubChem is maintained continuously. Schedule reruns when your workflow requires current source records.

Related Automation Lab Actors

• Open Food Facts EAN UPC GTIN Lookup for food-product barcode enrichment.
• PubMed Article Search Scraper for literature metadata and abstracts.
• NIST NVD CVE Vulnerability Records for another official public-data batch export workflow.

Choose this Actor when your primary entity is a PubChem compound and you need normalized chemistry fields rather than products, publications, or vulnerability records.