Loading ToolBench
Search books and authors, fetch editions, browse subjects, and resolve cover images from Open Library via MCP. STDIO or Streamable HTTP.
Strong tool definitions with consistent naming (verb_noun pattern), comprehensive descriptions (avg 180 chars), and complete JSON Schema with types and constraints. All 8 tools have input schemas with parameter descriptions. Pagination support (limit/offset) present across discovery tools. Output schemas are documented in descriptions but not formally in the schema JSON itself. Error handling guidance is present but could be more specific about recovery paths. Tool composition is excellent, each tool has a single responsibility and output IDs chain to downstream tools.
Fetch author detail by Open Library Author ID (OL…A). Returns bio, birth/death dates, photo IDs, and linked identifiers from Wikidata, VIAF, ISNI, Goodreads, and LibraryThing. Use openlibrary_search_authors to find an author ID first.
List works by an author. Returns titles, cover IDs, and work OLIDs for drilling into editions or details. Use openlibrary_get_author for author bio and details, or openlibrary_get_editions to explore specific printings.
Resolve a cover image URL for a book or author photo. Returns a direct HTTPS URL in the requested size (S/M/L). The Covers API always returns HTTP 200 — missing covers return a 1×1 placeholder GIF, not a 404 — so the identifier format is validated locally first: "id" must be numeric, "isbn" 10 or 13 digits, "olid" an edition OLID (OL…M) for target "book" and an author OLID (OL…A) for target "author". Identifiers with path separators or control characters, and author-by-ISBN lookups, are rejected before any request. URLs can be embedded in markdown as .
Resolve one or more editions by identifier: ISBN-10, ISBN-13, OCLC, LCCN, or Open Library Edition ID (OL…M). Every identifier in a call shares one id_type — pass id_type "isbn" for both ISBN-10 and ISBN-13. Up to 50 identifiers resolve in a single upstream request, so a bibliography or shelf export costs one call rather than one per book; a large batch is a large response, so ask for what you need. Returns full edition metadata including authors, publisher, language, all identifier types, and the parent work ID, with author names inline and no secondary lookup; when the edition record itself lists no authors, they are recovered from the parent work and marked as such. Partial success is the norm — identifiers that resolve come back in editions, the rest are listed in unresolved with a reason, and the call fails only when nothing resolved.
Output schemas not formally documented in JSON Schema format. Descriptions mention what is returned (e.g., 'Returns titles, cover IDs, and work OLIDs') but the actual response structure is not declared as a formal schema object. LLMs cannot reliably parse unstructured output descriptions.
Error handling descriptions are minimal. Tools state 'Partial success is the norm' (openlibrary_get_edition) but do not specify what error codes are returned, when to retry, or how to handle unresolved identifiers programmatically. Recovery guidance is implicit rather than explicit.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 70 | 2026-07-28+ | v2 |
List editions of a work — different publishers, languages, formats, and print runs. Returns ISBNs, publisher, language, page count, and edition OLIDs. Use after openlibrary_get_work or openlibrary_search_books to find a specific printing.
Browse works by subject. Returns matching works with edition counts and cover IDs, plus the total work count for the subject. Case and spacing are normalized before lookup, so "Science Fiction" and "science_fiction" are the same request. Subject tags are user-contributed and the vocabulary varies — when a subject returns no works, try a different word form (singular/plural), a synonym, or a broader term.
Fetch a work by Open Library Work ID (OL…W). Returns title, description, subjects, cover IDs, and linked author IDs for follow-up lookups. Works represent the abstract book concept independent of any specific edition. Note: author names are not included — use openlibrary_get_author or openlibrary_search_books for names.
Search Open Library authors by name. Returns Open Library Author IDs, names, birth/death dates, top works, and subject associations. Use author IDs for openlibrary_get_author (bio, remote IDs) or openlibrary_get_author_works (list of works).
No tool supports batch operations. openlibrary_get_edition accepts up to 50 identifiers in a single call (good), but other tools like openlibrary_get_author and openlibrary_get_work operate on single items only. Agents calling these in loops waste tokens and latency.
Parameter descriptions lack explicit format/constraint guidance. E.g., author_id states 'Open Library Author ID (OL…A)' but does not specify the regex pattern or character set. work_id similarly uses ellipsis notation instead of formal constraints. LLMs may hallucinate invalid IDs.
openlibrary_get_cover_url has complex validation logic (path separators, control characters, author-by-ISBN rejection) documented in the description but not enforced as schema constraints. The description is 380+ chars and mixes validation rules with usage guidance, reducing clarity.