---
title: PubChem exports and caching
description: Download CSV/SDF outputs and reuse cached PubChem results in the platform.
---

This page covers how to get [PubChem](/docs/pubchem) results as CSV, SDF, or JSON and how cached queries are reused. For a chat search, ask Ivy to save the results in the format you want; it writes the file in the chat's [code workspace](/docs/sandbox). A PubChem job submitted from the PubChem Database Search form has its own download on the job result. The fields you select most affect the outcome: exporting only what you need keeps files small and avoids the rate limits that large payloads can trigger.

## PubChem exports and caching

Export only the fields you need to keep files small and avoid rate limits.

## Export formats

- **CSV**: identifiers (CID, InChIKey), synonyms, formula, molecular weight,
  TPSA, HBA/HBD, rotatable bonds, logP estimates.
- **SDF**: structure plus selected properties for cheminformatics workflows.
- **JSON**: programmatic access to all returned fields when you need full
  fidelity.

## Selecting fields

- Tell Ivy which properties you need; PubChem can return large payloads.
- Default export includes identifiers and core physicochemical properties.

## Caching

- Recent queries may be cached to reduce round trips to PubChem.
- If cached, exports return immediately with the timestamp of the cached
  retrieval.
- Re-run the query if you need the latest data; the platform will refresh and update the
  cache.

## Good practices

- Keep query metadata (type, filters, timestamp) alongside exports for
  reproducibility.
- For large sets, export in pages rather than a single massive file.
- If you need assay links, ask Ivy for the PubChem bioassay data for those CIDs
  before exporting.

## Canvas visualization

When a PubChem search or job completes, a result card appears on the canvas as a compound data table with 2D structure depictions, following the platform's standard card shape — there's no PubChem-specific card size or border color.