| name | patent-landscape |
| description | Find out whether a chemical series is already claimed, using SureChEMBL's patent-extracted compound corpus and, where a key is available, PatentsView for legal status and assignee history. Use this skill to trace a structure to the patent documents that disclose it, survey an assignee's filings around a target, and understand what the freedom-to-operate question requires that a structure search cannot answer. Also trigger on SureChEMBL, patent chemistry, Markush structure, freedom to operate, composition of matter, assignee, priority date, patent family, or PatentsView. |
| license | MIT |
| allowed-tools | Read Write Edit Bash |
| compatibility | Requires Python 3.10+ and outbound HTTPS access to ftp.ebi.ac.uk. SureChEMBL has no public REST API — the bundled scripts navigate its bulk FTP tree, which is large, so plan for disk. PatentsView lookup is optional and needs a free PATENTSVIEW_API_KEY. Nothing here is legal advice; a freedom-to-operate opinion requires a qualified attorney. |
| metadata | {"version":"1.0","skill-author":"K-Dense Inc.","openclaw":{"emoji":"📜","homepage":"https://www.surechembl.org","envVars":[{"name":"PATENTSVIEW_API_KEY","required":false,"description":"Optional free PatentsView key for legal-status and assignee lookup."}]},"hermes":{"category":"research"}} |
Chemical Patent Landscape
Whether a series is already claimed decides whether it is worth pursuing, and the question comes
up long before anyone talks to an attorney. SureChEMBL extracts structures from patent full text,
images, and attachments — over 31 million compounds from the major offices — and it is the only
open, structure-searchable patent chemistry resource.
Services: https://ftp.ebi.ac.uk/pub/databases/chembl/SureChEMBL (bulk, no key) ·
https://search.patentsview.org/api/v1 (optional, free key).
Checked against: the live bulk tree, August 2026 — 31 releases, newest 2026-08-04.
Read references/surechembl-bulk.md before downloading anything,
references/patent-data-sources.md to know which source is
silent about what, and references/reading-patents.md before
drawing any conclusion — that one is judgement, not syntax, and the gap it describes is wide.
The two scripts
| Script | Answers |
|---|
surechembl_bulk.py | Which release, which tables, and how much disk? |
patent_search.py | Who is filing in this space, in the US? |
SureChEMBL has no REST API
Checked live: surechembl.org/api/ and every plausible variant return 404. The website is
interactive-only, and everything programmatic goes through the EBI bulk tree.
python skills/patent-landscape/scripts/surechembl_bulk.py releases --limit 3
# 31 releases, newest 2026-08-04. Updated fortnightly.
release url
2026-08-04 https://ftp.ebi.ac.uk/.../bulk_data/2026-08-04/
2026-07-17 ...
Two directories hold different data. bulk_data/ is SureChEMBL 2.0 — Parquet plus an FPSim2
similarity index, fortnightly. data/ is the legacy quarterly txt/SDF dump whose README is dated
2016. They are easily confused and the old one is much less useful.
Pin a release. Taking "latest" makes an analysis irreproducible against a corpus that changes
every two weeks.
Plan the download; it is 15 GB
python skills/patent-landscape/scripts/surechembl_bulk.py plan --question structure-to-patent
# structure-to-patent: match a structure to compound ids, then to the documents disclosing it
# 3 table(s), 14.1 GB for release 2026-08-04
curl -O https://ftp.ebi.ac.uk/.../compounds.parquet # 4.2 GB
curl -O https://ftp.ebi.ac.uk/.../patent_compound_map.parquet # 5.0 GB
curl -O https://ftp.ebi.ac.uk/.../patents.parquet # 5.9 GB
