What authentication does Knowhere API require?

Knowhere API may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.

Developer Tools

Publisher MCP server

Tools: 10

Knowhere API

The most accurate document parsing API for AI agents. Extract tables, formulas, and structured data with unmatched precision.

This listing identifies a third-party MCP server. MCPBundles is not affiliated with, sponsored by, or endorsed by the service or its operator unless stated on this page.

https://knowhereto.ai/

MCP

MCPBundles

Remote MCP URL

https://api.knowhereto.ai/mcp

This server may ask you to sign in when your AI client connects.

Last probed Nov 3, 2026 · api.knowhereto.ai

10tools discovered

For AI agents — MCPBundles Directory API
For MCP client apps (human connect):
MCP connect URL: https://api.knowhereto.ai/mcp
Skill page: https://www.mcpbundles.com/skills/knowhereto-mcp-0ed64faf92

Explore capabilities (10)

other 10

Other capabilities

Corpus.assets

other

corpus.assets

Find image and table chunks by type or query, or, given asset chunk_ids (host_of), list the sections that contain them. Rows show the containing section_path; an asset that no section contains is marked (not in any section). A plain listing does not tell you which assets matter: scope it to a section from a prior corpus.grep/corpus.recall hit, or use host_of on chunk_ids you already have.

Add to workspace

Tool inventory (10 tools)

Corpus.assets — Find image and table chunks by type or query, or, given asset chunk_ids (host_of), list the sections that contain them. Rows show the containing section_path; an asset that no section contains is marked (not in any section). A plain listing does not tell you which assets matter: scope it to a section from a prior corpus.grep/corpus.recall hit, or use host_of on chunk_ids you already have.
Corpus.grep — Exact string search over body text, image descriptions, table summaries and keywords, file names and section paths. Table cell values are not searched. Requires pattern or patterns — do not call without a term. Each term is a separate block of section rows (score-free — this is an exact match, not a ranked search; use corpus.recall for ranking). Terms with fewer matching sections are listed first. Rows past limit are folded per document with a reusable scope — copy that scope onto a later corpus
Corpus.list Documents — List every active document in the namespace with its summary, top keywords and chunk count. Use only when explicitly asked to inventory the namespace's documents, not as a first step for question answering.
Corpus.neighbors — Return documents related to the given document via the persisted document graph (typed-entity overlap, falling back to TF-IDF keyword overlap), along with the shared terms that justify each edge. Document-level only — there is no section- or entity-level graph yet.
Corpus.node Filter — Deterministic FOR-ALL/EXISTS/ANY/NOT filter over section titles and summaries — not body text (use corpus.grep for that). Narrows a map the same way corpus.outline does: the result is each matched section's full ancestor chain plus its entire descendant subtree (marked [Hit] where matched), not a bare list of hits. Call with scope plus a predicates array; each predicate is {field: 'path'|'summary', terms: [...], match: 'substring'|'regex'}. field=path matches the section path; field=summary matc
Corpus.outline — Map what a document (or one of its sections) covers: every section_path beneath a scope target, any depth, with title + summary, indented by level — no body text. Use this for 'what does this document/section cover' questions, not for finding an answer to a specific fact (use corpus.recall/corpus.grep for that). Accepts several scope targets in one call. Each target must be a non-leaf section (or the whole document, via a scope item with no section_path) — calling this on a section with no child

Get your MCP into directories

A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.

Directory coverage for brands

Frequently Asked Questions

What is the Knowhere API MCP server?

The most accurate document parsing API for AI agents. Extract tables, formulas, and structured data with unmatched precision.

How do I connect Knowhere API to my AI agent?

Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.

How many tools does Knowhere API provide?

Operate this server?

Request removal if you do not want Knowhere API listed here, or claim the listing if you want to manage it on MCPBundles.

Operate Knowhere API?

This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.

Claim this listing

MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms · Request removal

Corpus.read — Read full content for sections (document_id + section_path) or chunks (document_id + chunk_id) that you already located. Copy both values exactly from a prior result row. Page text shared with another section is filled in automatically. Images and tables the section contains are shown inline: small tables in full, large tables as row and column headers (then use corpus.query_table). Each ref's outcome (ok, or failed with a reason) is reported separately.

Corpus.recall — Fuzzy ranked candidate search for a question when you don't know where the answer lives. Ranks by how well section path and text match the query. Returns candidates with chunk_type, document_id, path and snippet. Table and image hits show their content; body hits also show the images and tables they contain. Exact string / identifier lookup is corpus.grep, not this tool. Read the candidate's document_id + section_path (or chunk_id, for image/table hits) with corpus.read next, not the filename.

Retrieval.query — Search published documents. Compose already happened in Knowhere. Returns grouped evidence as text and image content blocks, then a JSON text block with query, router_used, failure_reason, referenced_chunks, decision_trace, and results (raw path chunks for debug). Include navigation intent directly in your query text — the engine will automatically locate the right documents and sections.