@wyrdsekai/researchzosho-mcp
Deep research MCP server with citations and memory. Starts ResearchZosho for Claude Code, Codex, Gemini CLI or any MCP client that can run npx, and fetches the release if it is not installed.
npm · fingerprint e2378dea7fc1b2852705c1f5 · repository · RSS
What changed
Changed 2026-11-15 14:09 UTC
2 descriptions rewritten.
--- pinned/library_spend +++ observed/library_spend { - "description": "WHAT RESEARCH RUNS COST, at each model's list prices: the library's arithmetic on the servers' usage counts, not a bill. This month and in all, by host and by named model, with the cap where one is set, as hosts[] and entries[] and as text. Write access, as the settings.", + "description": "WHAT THE MODEL CALLS COST, at each model's list prices: the library's arithmetic on the servers' usage counts, not a bill. This month and in all, by host and by named model, with the cap where one is set, as hosts[] and entries[], and by what the calls were for (chat, nightly, research, translation, picture, other) as uses[]; and as text. Write access, as the settings.", "inputSchema": { "properties": { "patron": {
--- pinned/library_translate +++ observed/library_translate { - "description": "Translate a document the person has (txt, md, pdf, docx or epub) into a language and write it out in the same format or another, as a job. The document is cut into its units (headings, paragraphs, list items, table cells, captions) and each is translated in order with the library's glossary; the structure is kept, the layout is not. quality=checked (default) is one pass with the program's checks (every number, address, citation and glossary term on both sides, a length band, no unit dropped), the scores and a revision of what was flagged; quick is the one pass alone; reviewed and thorough add a second reading. What the checks still doubt after the revision is listed in the translation note, and the document holds the translation alone; notes=end numbers those passages and lists them at the end of the document too. Pass notes=end when the person asks for the doubts (the notes) at the end of the document and notes=off when they ask for none; without it the library's setting translate_notes decides, and the answer's notes says which the run was filed with. The answer is the plan: the units, the characters and the time; library_job follows the run, whose result names the file, the translation note beside it and both documents on the shelves. op=glossary lists the glossary, or with term and rendering adds to it. Paths are only accepted from the keeper (the terminal, the pages); raw names a document on the shelves.", + "description": "Translate a document the person has (txt, md, pdf, docx or epub) into a language and write it out in the same format or another, as a job. The document is cut into its units (headings, paragraphs, list items, table cells, captions) and each is translated in order with the library's glossary; the structure is kept, the layout is not. quality=checked (default) is one pass with the program's checks (every number, address, citation and glossary term on both sides, a length band, no unit dropped), the scores and a revision of what was flagged; quick is the one pass alone; reviewed and thorough add a second reading. What the checks still doubt after the revision is listed in the translation note, and the document holds the translation alone; notes=end numbers those passages and lists them at the end of the document too. Pass notes=end when the person asks for the doubts (the notes) at the end of the document and notes=off when they ask for none; without it the library's setting translate_notes decides, and the answer's notes says which the run was filed with. The answer is the plan: the units, the characters and the time, and translation: who translates, who reads and where the scores are made, with a sentence for the person in translation.said; library_job follows the run, whose result names the file, the translation note beside it and both documents on the shelves. op=glossary lists the glossary, or with term and rendering adds to it. A comic is translated too (manga, manhwa, manhua, a web comic's strips, a comic book, a graphic novel: a CBZ, a CBR, a ZIP or a folder of page pictures by itself; a PDF with comic=on): its pages are read by the comic helper on the library's machine, or on the machine that serves its heavy helpers, and the translation is set into the speech balloons. It is read from, and lettered in, Japanese, Chinese, Korean and the languages of the Latin, Cyrillic and Greek alphabets; from names the language of its lettering, and without it the helper finds the language on a sample of the pages (the answer's comic.from and comic.from_how say which language and how it is known; pass from when the person says it is another). A Japanese comic is read from the right, a Chinese one from the right when its lettering runs in columns, any other from the left: pass reads=right-to-left when the person says the comic reads from the right, as an English, French or Korean edition of a manga does (the answer's comic.reads and comic.reads_how say which way and whether it was said or found). The answer's comic.second_reading says whether the translator reads each balloon's lettering a second time (twice, for a comic that is not Japanese at checked and above when the translator sees pictures), or why the comic helper alone reads it (quick, blind, japanese), or at-start while that is to be found. Where the comic helper is not set up the call is answered with the error confirm, whose message says what a set-up fetches (about a gigabyte) and whose comic_helper lists each piece with its size and licence: show the person the message, and on their yes send the same call with setup=true. Paths are only accepted from the keeper (the terminal, the pages); raw names a document on the shelves.", "inputSchema": { "properties": { "as": { - "description": "The format written: md | txt | pdf | docx | epub (default: the input's).", + "description": "The format written: md | txt | pdf | docx | epub (default: the input's); for a comic cbz | pdf.", "enum": [ "md", "txt", - "epub" + "epub", + "cbz" ], "type": "string" }, + "comic": { + "description": "Whether the file is a comic: on (it is, whatever the program finds; a PDF of page pictures needs this) | off (it is a document). Default: an archive or a folder of page pictures is a comic by itself.", + "enum": [ + "on", + "off" + ], + "type": "string" + }, + "direction": { + "description": "The person's own words for the translation's voice, tone or wording, when they give one (\"as if you were Shakespeare\", \"it is a comedy, keep it whimsical\", \"mind the puns\"), at most 400 characters. The translator, the reviser and the reader are told it; numbers, dates, addresses, citations, names and glossary terms are kept and checked as in any run; a score or a length outside its band revises nothing, and the translation note lists where the translator could not keep it.", + "type": "string" + }, + "from": { + "description": "The language the document is written in, or a comic is lettered in: a name (Korean) or a code (ko). Default: what the program finds, from the document's text or, for a comic, on a sample of its pages.", + "type": "string" + }, "judge": { "description": "For checked: the model entry that reads source and translation side by side, or auto for the strongest other entry, local before hosted; default the library's judge seat, else the translator's model in fresh calls with a sample translated back and compared. A hosted entry is a paid call: the answer says its cost first (judge_cost_dollars).", "type": "string" + "reads": { + "description": "For a comic: the way its pages are read, when the person says it: right-to-left (an edition of a manga in any language) | left-to-right. Default: what the program finds, Japanese from the right, Chinese by its columns, any other language from the left.", + "enum": [ + "right-to-left", + "left-to-right" + ], + "type": "string" + }, "remove": { "description": "For op=glossary with term: remove the entry.", "type": "boolean" + "setup": { + "description": "The person's yes to setting the comic helper up, after a call answered with confirm: the run sets it up first, then translates the comic.", + "type": "boolean" + }, "term": { "description": "For op=glossary: the term as it stands in a source.", "type": "string"
Changed 2026-11-13 14:46 UTC
1 description rewritten.
--- pinned/library_translate +++ observed/library_translate { - "description": "Translate a document the person has (txt, md, pdf, docx or epub) into a language and write it out in the same format or another, as a job. The document is cut into its units (headings, paragraphs, list items, table cells, captions) and each is translated in order with the library's glossary; the structure is kept, the layout is not. quality=quick (default) is one pass; quality=checked adds the program's checks (every number, address, citation and glossary term on both sides, a length band, no unit dropped), a second reading of every unit by the judge's model when two are set, a revision of what was flagged, and a mark on what still stands. The answer is the plan: the units, the characters and the time; library_job follows the run, whose result names the file, the translation note beside it and both documents on the shelves. op=glossary lists the glossary, or with term and rendering adds to it. Paths are only accepted from the keeper (the terminal, the pages); raw names a document on the shelves.", + "description": "Translate a document the person has (txt, md, pdf, docx or epub) into a language and write it out in the same format or another, as a job. The document is cut into its units (headings, paragraphs, list items, table cells, captions) and each is translated in order with the library's glossary; the structure is kept, the layout is not. quality=checked (default) is one pass with the program's checks (every number, address, citation and glossary term on both sides, a length band, no unit dropped), the scores and a revision of what was flagged; quick is the one pass alone; reviewed and thorough add a second reading. What the checks still doubt after the revision is listed in the translation note, and the document holds the translation alone; notes=end numbers those passages and lists them at the end of the document too. Pass notes=end when the person asks for the doubts (the notes) at the end of the document and notes=off when they ask for none; without it the library's setting translate_notes decides, and the answer's notes says which the run was filed with. The answer is the plan: the units, the characters and the time; library_job follows the run, whose result names the file, the translation note beside it and both documents on the shelves. op=glossary lists the glossary, or with term and rendering adds to it. Paths are only accepted from the keeper (the terminal, the pages); raw names a document on the shelves.", "inputSchema": { "properties": { "as": { + "notes": { + "description": "Where what the checks still doubt is written: off (the document holds the translation alone; the translation note lists the doubts) | end (besides, each such passage numbered in brackets and listed at the end of the document). Default: the setting translate_notes, else off.", + "enum": [ + "off", + "end" + ], + "type": "string" + }, "op": { "description": "translate (default) | glossary.", "enum": [
Changed 2026-11-09 16:12 UTC
5 descriptions rewritten, 2 schema or annotation changes with the description left byte-identical, 20 tools added.
--- pinned/library_access +++ observed/library_access { - "description": "What became of an access request: pending, denied (with the reason), or approved, in which case the token, once, when the claim secret is right.", + "description": "What became of an access request (op=collect, the default, with request_id and claim): pending, denied (with the reason), or approved, in which case the token, once, when the claim secret is right. The keeper's side, for the person who keeps the library alone (anyone else is refused): op=readers lists who may read or write and the default; op=allow with did, level (deny, read or write) and name sets a reader's level; op=remove with did removes one; op=invite with name and level makes a link that signs a new person in once, shown once; op=requests lists who asked to be let in; op=approve with request_id and level, or op=deny with request_id and reason (or did), answers one; op=keys lists the keeper's keys for programs; op=key with label and level makes one, shown once; op=revoke with key_id revokes one. Nothing here is done unless the person asked for it in this conversation.", "inputSchema": { "properties": { "claim": { - "description": "The claim secret it returned.", + "description": "With collect: the claim secret it returned.", "type": "string" }, + "did": { + "description": "With allow, remove and deny: the reader's did.", + "type": "string" + }, + "key_id": { + "description": "With revoke: the key's id or prefix, as keys lists it.", + "type": "string" + }, + "label": { + "description": "With key: what the key is for.", + "type": "string" + }, + "level": { + "description": "With allow, invite, approve and key: deny | read | write.", + "enum": [ + "deny", + "read", + "write" + ], + "type": "string" + }, + "name": { + "description": "With allow: the reader's name; with invite: the new person's name, as they will sign in.", + "type": "string" + }, + "op": { + "description": "collect (default) | readers | allow | remove | invite | requests | approve | deny | keys | key | revoke.", + "enum": [ + "collect", + "readers", + "allow", + "remove", + "invite", + "requests", + "approve", + "deny", + "keys", + "key", + "revoke" + ], + "type": "string" + }, "patron": { "description": "Optional: who is asking ? {did, name, runtime}. Anonymous when absent.", "properties": { + "reason": { + "description": "With deny: why, in a sentence.", + "type": "string" + }, "request_id": { - "description": "The id library_request_access returned.", + "description": "With collect, approve and deny: the id library_request_access returned.", "type": "string" } }, - "required": [ - "request_id", - "claim" - ], + "required": [], "type": "object" }, "name": "library_access"
+ tool added: library_board
+ tool added: library_candidates
+ tool added: library_career
+ tool added: library_crews
+ tool added: library_export
+ tool added: library_follow
--- pinned/library_get +++ observed/library_get { "description": "One entry in full by id: body, every source with locator and edition, supersedes / superseded_by, subjects, related entries, and the review record {round, reviewer, decision, stale}. An investigation also carries sections[] {heading, chars}, sources[] as rows {n, locator, title, edition, published, fetched, language, same_as}, claims[] (its findings with bodies) and open_questions[]. A write-up can run past what one call may carry: pass section (a heading; \"answer\" = the write-up without the harness's sections; \"workers\", \"references\", \"evidence\") to get one section, and offset / max_chars for a window of the body.", "inputSchema": { + "in": { + "description": "Optional: a language (ja, English?); a report comes translated into it by the library's model, kept until the report changes, with written_in and read_in.", + "type": "string" + }, "max_chars": { "description": "Optional: at most this many characters of the body; chars and truncated say what was left.", "type": "integer"
+ tool added: library_home
+ tool added: library_images
--- pinned/library_inbox +++ observed/library_inbox { - "description": "The claims waiting for the keeper's decision ? drafts, and accepted claims whose review went stale ? and the decisions. op=list returns items[] {id, title, state, stale, kind, tier, confidence, writer, date, subjects, language, sources, report, report_title}, oldest first, with filters report (an investigation id or its prefix), subject, kind (extraction | synthesis | interpretation | speculation), tier, confidence, writer, state (draft | stale), language, q (words in the title). op=accept puts claims into every answer from now on; op=dispute files why; op=retire keeps them on disk and out of every answer ? each takes ids[] (or id), or report for every waiting claim of one investigation (write access).", + "description": "The claims waiting for the keeper's decision ? drafts, and accepted claims whose review went stale ? and the decisions. op=list returns items[] {id, title, state, stale, kind, tier, confidence, writer, date, subjects, language, sources, report, report_title}, oldest first, with filters report (an investigation id or its prefix), subject, kind (extraction | synthesis | interpretation | speculation), tier, confidence, writer, state (draft | stale), language, q (words in the title). op=accept puts claims into every answer from now on; op=dispute files why; op=retire keeps them on disk and out of every answer ? each takes ids[] (or id), or report for every waiting claim of one investigation (write access). op=list also returns reports[] {id, title, state, held_by, date, topic, reasons}: the reports that wait for the keeper's eye because one of their own checks stands against them (the write-up did not answer the question, or part of it; flagged sentences never sent for revision; most cited sources not read); op=accept with such a report's id (I-?) in id or ids accepts it as it stands.", "inputSchema": { "properties": { "confidence": {
--- pinned/library_job +++ observed/library_job { - "description": "One job from the ledger ({job}: state queued|running|done|failed|stopped, elapsed_s, result when finished), or your jobs (active[], finished[], paused) when job_id is omitted. op=stop {job_id} stops one run: queued, it never starts; running, it ends at its next turn. op=pause holds the runner (queued runs wait, a running one holds at its next turn); op=resume lets it go (write access).", + "description": "One job from the ledger ({job}: state queued|running|done|failed|stopped|offered, elapsed_s, result when finished), or your jobs (active[], finished[], paused) when job_id is omitted. op=stop {job_id} stops one run: queued, it never starts; running, it ends within seconds. op=pause holds the runner (queued runs wait, a running one holds at its next turn); op=resume lets it go (write access). op=choose {job_id, choice} answers a run held for one of its named choices, by number or by name; op=allow and op=decline answer for the person on a run the service held for their yes about help: a program sends any of the three only after it asked the person (write access).", "inputSchema": { "properties": { + "choice": { + "description": "With choose: the choice, by its number or its name as offered.", + "type": "string" + }, + "cursor": { + "description": "Without a job id: next_cursor from the previous page.", + "type": "string" + }, "job_id": { - "description": "The job to read or stop, e.g. J-0007.", + "description": "The job to read, stop or answer for, e.g. J-0007.", "type": "string" }, + "limit": { + "description": "Without a job id: how many finished runs a page holds (default 20).", + "type": "integer" + }, "op": { - "description": "read (default) | stop | pause | resume.", + "description": "read (default) | stop | pause | resume | choose | allow | decline.", "enum": [ "read", "stop", - "resume" + "resume", + "choose", + "allow", + "decline" ], "type": "string" },
+ tool added: library_look
+ tool added: library_mcp
--- pinned/library_models +++ observed/library_models { - "description": "The models this library can run on, by name: list them; add one (a name, a host address or word ? anthropic, openai, openrouter, gemini, nebius, huggingface ? a model id, prices per million tokens as in/out/cache, an effort, a monthly cap in dollars); remove one; keep a host's key. A key is never returned. Write access.", + "description": "The models this library can run on, by name: list them; add one (a name, a host address or word ? anthropic, openai, openrouter, gemini, nebius, huggingface ? a model id, prices per million tokens as in/out/cache, an effort, a monthly cap in dollars); remove one; keep a host's key; test one (whether its server answers now: answers, starting, or down, with a note when no key is kept for its host). The list states the default outright: default (name, model, host, drive) is the entry runs and chats that name no model use, summary says it in one sentence, and each entry's key is a word, kept, missing or not needed (a local server). A key is never returned. Write access.", "inputSchema": { "properties": { "cap": { - "description": "With add or remove: the entry's name ? letters, digits and hyphens.", + "description": "With add, remove or test: the entry's name ? letters, digits and hyphens.", "type": "string" }, "note": { - "description": "list (default) | add | remove | key", + "description": "list (default) | add | remove | key | test", "enum": [ "list", "add", - "key" + "key", + "test" ], "type": "string" },
+ tool added: library_patent_abstracts
+ tool added: library_patent_claims
+ tool added: library_patents
+ tool added: library_records
+ tool added: library_repo
--- pinned/library_research +++ observed/library_research { "description": "A RESEARCH RUN: send a question to The Librarian's workers. The library's own runner plans sub-questions (yours, if you pass them), researches each in parallel with web search + fetch noting every fact with its source, has a critic decide on a second round, and writes the investigation in sections; the result enters the library as a draft investigation attributed to you and the review crew extracts findings from it. Returns a job_id to poll with library_job. Write access. Takes minutes to hours ? do not wait inline. A field that acts only when asked for, such as family history, joins the run only when field names it; a question that looks like one of its questions is researched as ordinary research, and the result then carries suggestion {field, why, how}: what the field's mode does and how to ask for it.", "inputSchema": { + "citations": { + "description": "Optional: how the report's sources are written: numbered (the default), apa, mla, chicago or ieee. The numbers in the text stay, so every cited sentence is still checked. Without it, the library's setting citation_style. Since 0.7.0.", + "enum": [ + "numbered", + "apa", + "mla", + "chicago", + "ieee" + ], + "type": "string" + }, "collections": { "description": "Optional: names of the person's collections (folders shelved with `researchzosho add <dir>`) to search.", "items": { + "goal": { + "description": "Optional: the goal the run serves, by name (home, work, dossier, buy, prior-art, family, choose, learn-a-craft, generic), which chooses the report's shape and the instruments the run brings; a field's name is taken as its goal until the goals land. Since 0.7.0.", + "type": "string" + }, + "length": { + "description": "Optional: how long the report is: short (about 400 words), medium (about 1,200), long (about 3,000), or a number of words. Without it, the library's setting report_length, else as long as the question needs. Since 0.7.0.", + "type": "string" + }, "locale": { "description": "Optional: the person's language and country as a language tag, such as en-US, es-MX or ja-JP. When a question reads as a person asking about harming themselves, where to find help comes in that language (English, Spanish or Japanese; English otherwise) with that country's services first, in the message and as data (error.data.helplines, or help.helplines beside a run).", "type": "string" + "paid": { + "description": "Optional: false leaves the paid search places and the paid page reader off for this run (Exa, Tavily, the keeper's own search service, Serper, SerpApi, Brave, Firecrawl); the free places answer. True unless given. Since 0.7.0.", + "type": "boolean" + }, "patron": { "description": "Optional: who is asking ? {did, name, runtime}. Anonymous when absent.", "properties": { + "shape": { + "description": "Optional: report (the default); reading-list, an annotated bibliography, each source with what it is, why it matters for the question, how strong it is and its link; outline, headings with one line each. A question that asks for a format in its own words gets that format. Since 0.7.0.", + "enum": [ + "report", + "reading-list", + "outline" + ], + "type": "string" + }, "sources": { "description": "both (the shelves first, then the web ? default) | shelves (the person's own corpus only, no web) | web.", "enum": [
--- pinned/library_settings +++ observed/library_settings { - "description": "The library's settings (write access): op=list shows every setting with what it is for and its value, a key only as set or not; op=set with name and value writes one (name as the key, or without its RESEARCHZOSHO_ prefix; a blank value, or op=unset, unsets it). A program sets a key or an address only when the person gave it in this conversation.", + "description": "The library's settings (write access): op=list shows every setting with what it is for and its value, a key only as set or not; op=set with name and value writes one (name as the key, or without its RESEARCHZOSHO_ prefix; a blank value, or op=unset, unsets it). A program sets a key or an address only when the person gave it in this conversation. op=read_in with value sets the language the person reads reports in (a blank unsets it); read access is enough. op=check asks every outside service whose key is set whether it is accepted and what is left, keeps the answers and returns checks[].", "inputSchema": { "properties": { "name": { - "description": "list (default) | set | unset.", + "description": "list (default) | set | unset | read_in | check.", "enum": [ "list", "set", - "unset" + "unset", + "read_in", + "check" ], "type": "string" },
+ tool added: library_sift
+ tool added: library_sources
+ tool added: library_spend
+ tool added: library_translate
+ tool added: library_video
Changed 2026-11-02 15:50 UTC
1 schema or annotation change with the description left byte-identical, 5 tools added.
+ tool added: library_dossier
+ tool added: library_help
--- pinned/library_research +++ observed/library_research { "description": "A RESEARCH RUN: send a question to The Librarian's workers. The library's own runner plans sub-questions (yours, if you pass them), researches each in parallel with web search + fetch noting every fact with its source, has a critic decide on a second round, and writes the investigation in sections; the result enters the library as a draft investigation attributed to you and the review crew extracts findings from it. Returns a job_id to poll with library_job. Write access. Takes minutes to hours ? do not wait inline. A field that acts only when asked for, such as family history, joins the run only when field names it; a question that looks like one of its questions is researched as ordinary research, and the result then carries suggestion {field, why, how}: what the field's mode does and how to ask for it.", "inputSchema": { + "seek": { + "description": "Optional: a seek group's name (letters, digits, dashes). The run reads and writes its own copy of the shelves under that name, so several seeks of one question never see each other's work; library_seeks op=merge writes one report from them.", + "type": "string" + }, "sources": { "description": "both (the shelves first, then the web ? default) | shelves (the person's own corpus only, no web) | web.", "enum": [
+ tool added: library_seeks
+ tool added: library_suggest
+ tool added: library_tell
Changed 2026-10-18 11:02 UTC
1 schema or annotation change with the description left byte-identical, 1 tool added.
+ tool added: library_models
--- pinned/library_research +++ observed/library_research { "description": "A RESEARCH RUN: send a question to The Librarian's workers. The library's own runner plans sub-questions (yours, if you pass them), researches each in parallel with web search + fetch noting every fact with its source, has a critic decide on a second round, and writes the investigation in sections; the result enters the library as a draft investigation attributed to you and the review crew extracts findings from it. Returns a job_id to poll with library_job. Write access. Takes minutes to hours ? do not wait inline. A field that acts only when asked for, such as family history, joins the run only when field names it; a question that looks like one of its questions is researched as ordinary research, and the result then carries suggestion {field, why, how}: what the field's mode does and how to ask for it.", "inputSchema": { + "model": { + "description": "Optional: the named model to run on, one of the library's model entries by name (library_models op=list names them), or auto for best judgement, which is the default when more than one is named. Each entry is a host, a model id and its prices; the run's report says which ran it and what it cost.", + "type": "string" + }, "patron": { "description": "Optional: who is asking ? {did, name, runtime}. Anonymous when absent.", "properties": {
Changed 2026-10-16 11:25 UTC
1 tool added.
+ tool added: library_share
Changed 2026-10-13 14:14 UTC
5 tools added.
+ tool added: library_dna
+ tool added: library_my_shelf
+ tool added: library_owned
+ tool added: library_prefs
+ tool added: library_react
Changed 2026-10-09 13:15 UTC
1 tool added.
+ tool added: library_settings
Current tools
Show all 69 tool fingerprints
Watch this server yourself
If you run this server, put the proxy in front of it. It pins these exact fingerprints on first connect and stops the session if they move.
npx -y mcp-pin@0.2.4 -- <your @wyrdsekai/researchzosho-mcp command>
Or subscribe to this page's RSS feed to be told when it changes.
Badge
The badge states one fact about time and nothing else. It never claims a server is safe.
[](https://mcp-pin.gautamkhosla.com/servers/8e31a6bdb76a908c.html)