# Tools reference Complete reference of every tool the GroupDocs.Search MCP server exposes to AI agents, with parameters, example prompts, and results. Complete reference of every tool the GroupDocs.Search MCP server exposes to AI agents, with parameters, example prompts, and results. Captured from a live `tools/list` call against server version **26.10.1** (raw capture: `tools-list.generated.json` in this section's source). | Tool | What it does | |---|---| | [`add_to_index`]({{< ref "search/mcp/tools-reference/add-to-index.md" >}}) | Indexes a file or folder from the storage folder into a named index (created if missing) | | [`search`]({{< ref "search/mcp/tools-reference/search.md" >}}) | Finds the documents that match a query — words, phrases, boolean, fuzzy, date and number ranges | | [`get_text_fragments`]({{< ref "search/mcp/tools-reference/get-text-fragments.md" >}}) | Returns the matching passages with surrounding words, per document | | [`update_index`]({{< ref "search/mcp/tools-reference/update-index.md" >}}) | Re-reads changed, added and deleted files in an index's paths | | [`remove_from_index`]({{< ref "search/mcp/tools-reference/remove-from-index.md" >}}) | Removes a path that was added with add_to_index | | [`get_index_info`]({{< ref "search/mcp/tools-reference/get-index-info.md" >}}) | Reports an index's documents, paths, status and the indexes that exist | | [`create_index`]({{< ref "search/mcp/tools-reference/create-index.md" >}}) | Creates an empty named index, or empties one with recreate | | [`get_license_status`]({{< ref "search/mcp/tools-reference/get-license-status.md" >}}) | Reports the active licensing mode and, under metered licensing, consumption | ## How the tools address documents and indexes No tool takes a document's content. They take **paths inside the storage folder** and **index names**: | Argument | Shape | Example | |---|---|---| | `documentsPath` / `documentPath` | relative to the storage folder (`/data` in Docker) — a file or a folder; omit to mean the whole folder | `"contracts"`, `"contracts/lease.pdf"` | | `indexName` | a name, never a path; omit for the index called `default` | `"contracts"`, `"legal-2026"` | | `query` | the search language below | `"termination AND penalty"` | Indexes live in the index store — `/data/.groupdocs-search-index/` by default, or under `GROUPDOCS_MCP_INDEX_PATH` when set — and **persist between calls and restarts**. The server never indexes its own index folder. Result paths come back relative to the storage folder (`contracts/acme-master-services-agreement.pdf`); a match inside an email or archive adds `innerPath` (the attachment's name). ## The query language | You want | Query | Note | |---|---|---| | a word | `penalty` | case-insensitive unless `caseSensitive: true` | | a phrase | `"termination penalty"` | **always quote multi-word phrases** — unquoted words do not match | | all of / any of / not | `termination AND penalty` · `renewal OR renews` · `lease NOT draft` | also `&`, `\|`, `!`; group with parentheses | | one unknown character | `terminat?` | matches `terminate`, not `termination` — `?` is one character | | spelling variants | `renewl` with `fuzzy: true` | also matches near words — check the `terms` it found | | dates written in documents | `content: daterange(2026-07-01 ~~ 2026-09-30)` | **use `content:`** — see the note below | | numbers in a range | `content: (100000 ~~ 999999)` | whole numbers written without separators | | a file-name word | `filename: q3` | matches words in the file name | | a regular expression | `^inv[0-9]+` | a query starting with `^` | There is **no trailing `*` wildcard**: use `?` patterns or `fuzzy: true` for variants. `AND` is evaluated per **document**, not per sentence — confirm with [`get_text_fragments`]({{< ref "search/mcp/tools-reference/get-text-fragments.md" >}}) that the words are about the same thing. {{< alert style="warning" >}} **Put `content:` in front of a date range.** The engine also indexes each file's creation and modification dates as fields, so a bare `daterange(…)` matches every file written in that period. `content: daterange(…)` searches only the dates written inside the documents. To search by file date on purpose, use `modificationdate: daterange(…)`. {{< /alert >}} The full grammar: [query language specification]({{< ref "search/net/developer-guide/advanced-usage/searching/query-language-specification.md" >}}). ## Contents - [add_to_index](https://docs.groupdocs.com/search/mcp/tools-reference/add-to-index.md) - [search](https://docs.groupdocs.com/search/mcp/tools-reference/search.md) - [get_text_fragments](https://docs.groupdocs.com/search/mcp/tools-reference/get-text-fragments.md) - [update_index](https://docs.groupdocs.com/search/mcp/tools-reference/update-index.md) - [remove_from_index](https://docs.groupdocs.com/search/mcp/tools-reference/remove-from-index.md) - [get_index_info](https://docs.groupdocs.com/search/mcp/tools-reference/get-index-info.md) - [create_index](https://docs.groupdocs.com/search/mcp/tools-reference/create-index.md) - [get_license_status](https://docs.groupdocs.com/search/mcp/tools-reference/get-license-status.md)