paperlib-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SEARXNG_URL | No | Optional. A SearXNG instance with JSON output enabled, used as a last-resort search fallback (never a download source). | |
| CORE_API_KEY | No | Optional; raises CORE's limits. | |
| NCBI_API_KEY | No | Optional; raises PubMed from 3 to 10 requests/s. | |
| PAPERLIB_CACHE | No | Read-through cache for the S3 store. Default: ~/.cache/paperlib. | |
| PAPERLIB_EMAIL | No | Contact email. Sent to OpenAlex (polite pool), NCBI and PubMed, and in the User-Agent. Required by Unpaywall; without it the unpaywall route is skipped. Never sent unless set. | |
| PAPERLIB_STORE | No | Store type: 'local' (default), 's3', or 'memory'. | |
| OPENALEX_API_KEY | No | Optional (free). Raises limits and enables OpenAlex's own cached full texts. A wrong key fails loudly rather than silently degrading. | |
| PAPERLIB_DATA_DIR | No | Local store directory. Default: ~/.local/share/paperlib ($XDG_DATA_HOME/paperlib). | |
| PAPERLIB_PROVIDERS | No | Comma-separated locate order (default below). | |
| PAPERLIB_S3_BUCKET | No | S3 store: bucket name (required when PAPERLIB_STORE=s3). | |
| PAPERLIB_S3_REGION | No | S3 store: region. | |
| PAPERLIB_WEB_FALLBACK | No | Set to '0' to disable the web fallback. | |
| PAPERLIB_S3_PUBLIC_URL | No | S3 store: anonymous base URL used to prove objects are not public. Default: https://<bucket>.<endpoint host>. | |
| PAPERLIB_S3_ENDPOINT_URL | No | S3 store: endpoint URL for non-AWS services (DigitalOcean Spaces, MinIO, R2, etc.). | |
| PAPERLIB_S3_ACCESS_KEY_ID | No | S3 store access key ID; if unset, boto3's usual credential chain applies. | |
| PAPERLIB_SEARCH_PROVIDERS | No | Comma-separated federated-search set (default below). | |
| OPENCITATIONS_ACCESS_TOKEN | No | Optional; OpenCitations asks applications to send one. | |
| PAPERLIB_S3_SECRET_ACCESS_KEY | No | S3 store secret access key; if unset, boto3's usual credential chain applies. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| searchB | Search all configured literature providers and mark papers already held. |
| fetchB | Fetch and privately store a legal open-access copy of a paper (held papers cost nothing). |
| libraryA | Search papers already held (titles and authors; optionally inside stored full text). |
| textC | Read a page of a held paper's machine-extracted full text. |
| provenanceA | Return acquisition route, licence, checksums, and every route tried for a held paper. |
| citationsB | Works citing a paper (direction='citations') or cited by it ('references'), via OpenCitations. Identifiers only, and no citation intent: a citation is not agreement. |
| providersA | List every provider, what it can do, what it needs, and its terms. No network. |
| statusA | Counts of works held, with and without full text, and OpenAlex allowance remaining. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 8 tools
Each tool targets a distinct aspect of the paper workflow: search for external discovery, library for held papers, fetch for acquisition, text for reading, provenance for metadata, citations for references, providers for source configuration, and status for aggregate counts. There is no overlap or ambiguity in their purposes.
All tool names are single lowercase words, forming a consistent and predictable pattern. Although the names mix verbs (fetch, search) and nouns (library, status), the uniform lowercase style establishes a coherent convention that is easy to follow.
With 8 tools, the server is well-scoped for a paper management and citation service. Each tool covers a necessary function without redundancy, and the count feels neither sparse nor overwhelming.
The tool set covers the full lifecycle of paper handling: discover (search), acquire (fetch), search held (library), read (text), inspect provenance, explore citations, understand providers, and check status. No obvious gaps exist for the stated domain—every action a user would reasonably need is represented.