Navigate 200 Million Published Articles with BioMCP¶
BioMCP federates PubMed, Semantic Scholar, Europe PMC, and PubTator3 into a single article CLI. Search, read, triage, traverse citation graphs, extract entities, and compare papers — all without leaving the terminal. Nine capabilities, one grammar.
1. Search¶
Find papers by gene, disease, drug, author, keyword, date range, journal, article type, or any combination. Results are deduplicated across sources and ranked by relevance.
$ biomcp search article --gene BRAF -d melanoma --limit 5
# Articles: gene=BRAF, disease=melanoma, sort=relevance
| PMID | Title | Source(s) | Date | Cit. |
|----------|----------------------------------------------------|-------------------|------|------|
| 30544808 | Oleuropein, the Main Polyphenol of Olea europaea… | Semantic Scholar | 2018 | 100 |
| 29295999 | ERK-mediated phosphorylation regulates SOX10… | Semantic Scholar | 2018 | 70 |
| 26790143 | p53 Reactivation by PRIMA-1(Met) sensitises… | Semantic Scholar | 2016 | 52 |
| 28455392 | Biomarker Accessible and Chemically Addressable… | Semantic Scholar | 2017 | 45 |
You can stack filters: --gene BRAF --drug vemurafenib --since 2020 --type review --sort citations. Twelve filter options, all combinable.
2. Article Detail¶
Get the abstract, authors, DOI, citation count, open-access status, and suggested follow-up commands in one call.
$ biomcp get article 12068308
# Mutations of the BRAF gene in human cancer.
PMID: 12068308
DOI: 10.1038/nature00766
Journal: Nature
Date: 2002-06-27
Citations: 7469
Type: research support, non-u.s. gov't
Open Access: No
BioMCP auto-generates follow-up commands for every gene, drug, disease, and mutation mentioned in the abstract. You don't have to know the identifiers — they're extracted for you.
3. AI Summary (TLDR)¶
Semantic Scholar generates a one-sentence machine summary for most papers. Instead of reading the full abstract, get the TLDR.
$ biomcp get article 12068308 tldr
TLDR: BRAF somatic missense mutations in 66% of malignant melanomas
and at lower frequency in a wide range of human cancers, with a single
substitution (V599E) accounting for 80%.
Citations: 10,599
Influential citations: 532
References: 28
Two citation counts: PubMed's 7,469 (MEDLINE-indexed) vs Semantic Scholar's 10,599 (includes preprints and conference papers). The "influential citations" count (532) is Semantic Scholar's metric for citations that meaningfully shaped the citing work.
4. Citations¶
Who cited this paper? Semantic Scholar returns forward citations with context snippets, intent labels, and influence markers.
$ biomcp article citations 12068308 --limit 3
# Citations for PMID 12068308
| PMID | Title | Intents | Influential |
|------|----------------------------------------------------|---------|-------------|
| … | FDG PET/CT as the Decisive Modality in… | - | no |
| … | Associations of Tumor Somatic Mutations and… | - | no |
| … | A Complete Response to Immunotherapy in a… | - | no |
When available, each citation includes a context snippet (the sentence where your paper was cited) and an intent label (Background, Method, or Result). This tells you why someone cited a paper, not just that they did.
5. References¶
What did this paper cite? Trace the foundational work backward.
Same structure as citations but in the opposite direction. Follow the intellectual lineage to find the papers that a given work builds on. Useful for literature reviews: start with a recent review, then walk backward through its reference list.
6. Similar Papers¶
Content-based recommendations from Semantic Scholar. Finds related work by meaning, not just shared keywords.
$ biomcp article recommendations 12068308 --limit 3
# Recommendations for PMID 12068308
| PMID | Title | Journal | Year |
|------|----------------------------------------------------|-----------------|------|
| … | Oncogenes and tumor suppressor genes enriched in… | bioRxiv | 2026 |
| … | Genome-wide CRISPR Screen Identifies Menin as… | Cancer Research | 2026 |
| … | Pancreas cancer: Genomics, RAS therapeutics… | Cancer Research | 2026 |
Multi-seed with negative filtering: biomcp article recommendations 12068308 21639808 --negative 39073865 finds papers similar to the first two but NOT similar to the third. Useful for narrowing a broad literature search.
7. Entity Commands¶
BioMCP extracts every gene, disease, drug, and mutation from a paper and turns them into runnable commands. This is the bridge between articles and everything else in BioMCP.
$ biomcp article entities 12068308
## Genes (3)
| Entity | Mentions | Next Command |
|--------|----------|---------------------------|
| BRAF | 4 | biomcp get gene BRAF |
| MEK | 1 | biomcp get gene MEK |
| RAF | 1 | biomcp get gene RAF |
## Diseases (4)
| Entity | Mentions | Next Command |
|--------------------|----------|-------------------------------------------|
| cancer | 5 | biomcp search disease --query cancer |
| malignant melanoma | 1 | biomcp search disease --query "malignant… |
From any paper, jump to any entity. Get the gene's pathways, the drug's interactions, the disease's clinical trials — without looking up identifiers yourself.
8. Full Text¶
Retrieve open-access full text when available. Semantic Scholar provides PDF links with license info.
$ biomcp get article PMC9984800 fulltext
# Precision oncology for BRAF-mutant cancers… - fulltext
## Full Text (PMC OA)
Saved to: /tmp/biomcp/b379bd903115c5acbb663120e8bed420.txt
Full text is retrieved from PubMed Central's Open Access subset. Not all papers are available — when they're not, BioMCP tells you and suggests alternatives (TLDR, annotations).
9. Batch Compare¶
Compare multiple papers side by side: TLDRs, citation counts, influence metrics, and extracted entities for your shortlist in one command.
$ biomcp article batch 12068308 21639808 22663011
## 1. Mutations of the BRAF gene in human cancer.
PMID: 12068308 | Nature | 2002
TLDR: BRAF somatic missense mutations in 66% of malignant melanomas,
V599E accounting for 80%.
Citations: 10,599 (influential: 532)
## 2. Improved survival with vemurafenib in melanoma with BRAF V600E mutation.
PMID: 21639808 | N Engl J Med | 2011
TLDR: Vemurafenib produced improved rates of overall and progression-free
survival in a phase 3 randomized clinical trial.
Citations: 6,756 (influential: 175)
Pass any number of PMIDs, PMCIDs, or DOIs. Useful for triage: run a search, pick your candidates, then batch them to compare TLDRs and citation metrics before deciding which papers to read in full.
API Keys¶
BioMCP works out of the box with no API keys. All nine article capabilities are functional immediately after install. Two optional keys improve reliability:
Semantic Scholar (S2_API_KEY): Free key from semanticscholar.org/product/api. Without it, you share a rate limit with all unauthenticated users (1 request every 2 seconds). With the key, you get a dedicated rate limit (1 request per second). Citations, references, recommendations, TLDRs, and batch enrichment all benefit.
NCBI (NCBI_API_KEY): Free key from ncbiinsights.ncbi.nlm.nih.gov. PubTator3, PubMed, and PMC Open Access all use NCBI infrastructure. Without a key, you're limited to 3 requests per second. With the key, 10 requests per second.
Check your configuration with biomcp health --apis-only — it shows which APIs are authenticated and which are using shared pools.
Try it¶
Then start with a paper you know: