Zoteus

Save structured fields extracted from papers

zotero_save_extraction

DestructiveIdempotent

Saves structured paper extractions as tagged child notes in Zotero, storing named fields, values, source quotes, and verification verdicts so reviews stay readable and machine-readable.

Instructions

Save a structured extraction (named fields with values, such as sample_size, method, country, outcome) for one or many papers, as a child note on each paper: a readable table plus a machine-readable copy, tagged "zoteus-extraction". Zotero has no custom fields; a note syncs and anyone can read it in Zotero. YOU read the paper (zotero_get_fulltext) and supply the values: this server has no language model. Give each value the verbatim quote it rests on, and its page: when the paper's full text is available (Zotero's index, or the PDF or EPUB itself) the quote is checked against it and the verdict ("verified", "not-found" or "unchecked", with the page it was found on when known) is stored beside the value. A quote that was not found is still saved, and named in the result, so you can correct it. schema names the extraction (e.g. "rct-review"): one note per paper and schema. Saving again for the same paper and schema UPDATES that note in place: fields you send replace fields of the same name, a null value removes a field, fields you do not mention are kept (pass replace_all:true to drop them), and the result lists every replaced and removed value with what it was. Nothing else in the library is touched. Read the extractions back as one table with zotero_extraction_table. Writes go to the running Zotero desktop app (Zotero 10+) for every library it holds, otherwise to the cloud Web API.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "schema": "<schema>",
      "items": "<items>"
    },
    "name": "zotero_save_extraction"
  }
}

Input Schema

Name	Required	Description
`items`	Yes	One entry per paper, at most 50.
`schema`	Yes	Name of the extraction schema, e.g. "rct-review". Papers saved under the same name form one table.
`library_id`	No	Numeric id of the library to address, e.g. 5234875 for a group (zotero_groups lists the ids you can reach). Omit to use the configured default library; an id given without library_type is read as a group id.
`replace_all`	No	Drop fields already saved for a paper that this call does not mention (default false: they are kept).
`library_type`	No	Which library to address: "user" (a personal library) or "group" (a shared group library). Omit to use the library this server is configured for. "group" on its own is refused: pass library_id with it.
`verify_quotes`	No	Check each quote against the paper's full text (default true). Off skips reading the text and stores quotes as "unchecked".

Output Schema

Name	Required	Description
`items`	Yes	One entry per paper, in the order given.
`saved`	Yes	Papers whose note was created or updated.
`failed`	Yes	Papers that failed.
`schema`	Yes	The schema saved under.
`target`	No	Where the write went: "local" (Zotero desktop local API), "desktop" (connector protocol) or "cloud" (Zotero Web API).

Schema Changelog

Changes observed during successful MCP inspections.

Addedv1.25.1Nov 10, 2026

TDQS

A4.8/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well beyond the annotations: it discloses that notes are used because Zotero has no custom fields, that quotes are checked against full text with three verdicts stored beside the value, that unverified quotes are still saved and named in the result, that saves are upserts per (paper, schema), and that nothing else in the library is touched. The mutation and idempotency claims match destructiveHint/idempotentHint rather than contradicting them.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the core action and storage target, then progresses through verification, update semantics, and read-back in a logical order. It is a single dense paragraph that runs long, and a couple of clauses (the Zotero desktop/Web API routing note) could be trimmed, but almost every sentence carries information.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a mutation tool with a rich output schema, it covers what an agent needs before calling: required inputs, verification behavior, upsert/removal semantics, and where writes land. Return-value details are partly deferred to the output schema, which is appropriate.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3; the description still adds useful meaning on top by explaining the merge semantics of a null value, the kept-field default, and that `schema` groups papers into one table. It does not add much beyond the schema for library_id/library_type, which is acceptable given the schema already documents them.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource (save a structured extraction) and pins down the storage target (a child note per paper, tagged "zoteus-extraction"). It explicitly separates its role from zotero_get_fulltext ("YOU read the paper") and zotero_extraction_table ("Read the extractions back as one table"), so it is distinguishable from siblings without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use routing: call zotero_get_fulltext first, supply values yourself, and use zotero_extraction_table to read back. It also states the when-not-to-keep existing fields condition and the replace_all override, plus the fallback of desktop app vs Web API write routing.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server