MCP Endpoints
Valency MCP provides 32 endpoints organized into 8 categories. Once connected, your AI assistant can use any of these tools. Try some example prompts.
Search & Discovery
Section titled “Search & Discovery”| Endpoint | Description |
|---|---|
semantic_search_papers | Search by meaning using semantic embeddings |
search_by_title | Find papers by title keywords |
search_by_abstract | Find papers by abstract content |
search_by_author | Find papers by author name (literal string matching) |
search_by_category | Find papers in a research domain (recency-ordered) |
search_by_venue | Search by publication venue (journal or conference) |
search_cross_category | Search abstracts across multiple categories |
find_similar_papers | Find semantically similar papers to a given paper |
get_citing_papers | Find papers that cite a given paper |
Filtering
Section titled “Filtering”| Endpoint | Description |
|---|---|
filter_by_categories | Filter across multiple domains (recency-ordered) |
filter_by_date_range | Filter by publication date |
filter_by_license | Filter by license type (CC-BY, CC0, arxiv default, etc.) |
filter_papers_with_doi | Find papers with DOI identifiers |
count_papers | Count papers matching filters (number only) |
Author Intelligence
Section titled “Author Intelligence”| Endpoint | Description |
|---|---|
get_author_profile | Comprehensive researcher summary profile |
find_coauthors | Find an author’s most frequent collaborators |
compare_authors | Compare 2–10 researchers side by side |
batch_author_categories | Category distributions for multiple authors |
identify_prolific_authors | Rank authors by publication count |
Paper Details
Section titled “Paper Details”| Endpoint | Description |
|---|---|
get_paper_by_id | Direct lookup by paper ID |
get_paper_versions | Revision history for a paper |
Trends & Analytics
Section titled “Trends & Analytics”| Endpoint | Description |
|---|---|
get_keyword_trends | Paper counts over time for a keyword/abstract search |
get_publication_trends | Paper counts over time for a category |
get_publication_trends_batch | Trends for multiple categories in one call |
analyze_corpus_metrics | Corpus-wide descriptive statistics |
identify_research_domains | Rank research domains by volume |
get_field_coverage | Most common values for a metadata field |
Corpus Info
Section titled “Corpus Info”| Endpoint | Description |
|---|---|
list_sources | Available paper sources with counts and date ranges |
Export
Section titled “Export”| Endpoint | Description |
|---|---|
export_papers_csv | Export papers to CSV |
export_papers_json | Export papers to JSON |
export_from_filter | Export papers matching a filter as in-memory content |
Feedback
Section titled “Feedback”| Endpoint | Description |
|---|---|
submit_feedback | Submit feedback about search quality, bugs, or feature requests |
Getting Started with Endpoints
Section titled “Getting Started with Endpoints”Once you’ve connected Valency to your AI assistant, you can start using these endpoints immediately through natural language.
Example queries
Section titled “Example queries”Semantic search:
“Search for papers about transformer attention mechanisms”
Author research:
“Show me Geoffrey Hinton’s publication profile” “Who are Yann LeCun’s most frequent collaborators?”
Trends:
“What are the publication trends for machine learning over the last 5 years?”
Filtering:
“Find papers about climate change from 2023 with CC-BY license”
Export:
“Export the top 100 papers about CRISPR to CSV”
- Be specific: More specific queries return better results
- Combine endpoints: Your AI assistant can chain multiple endpoints together
- Use filters: Narrow results by date, license, or category
- Try semantic search first:
semantic_search_papersoften gives the best results for exploratory queries