Semantic Search & RAG Tool

The Google Drive Semantic Search Tool

Search your vectorized Google Drive by meaning across files and folders, filtered by MIME type, folder, and date — the retrieval layer for any agent that reasons over your documents.

Explore VDF AI Agents
MeaningSemantic, not filename, matching
FilteredType, folder, and date
ScoredEvery hit ranked by similarity
100%On-prem, files never leave
The Shared Drive Problem

The document exists — buried three folders deep

Drive search matches filenames and full text literally. When files are named "final_v3" and scattered across nested shared folders, the document you need is nearly impossible to find by name.

01

Filenames lie

"Untitled document" and "final_v3_REAL" tell you nothing about what a file contains.

02

Folders hide files

The right document sits in a shared drive or nested folder you didn’t think to open.

03

No relevance ranking

Native search floods you with weak matches and no confidence signal.

04

Files can’t leave

Contracts, designs, and plans are sensitive — they can’t go to a hosted search service.

How the Tool Works

Meaning-aware search over your Drive

Semantics

Describe it, don’t name it

Find the file by what it contains.

The tool embeds your query and matches it against vectorized Drive files, surfacing the right document even when its filename is meaningless or shares no keywords with your question.

  • Content-based matching
  • Across files and folders
  • Similarity score per hit
  • Tunable top_k up to 50
Intent
Meaning Match

Beyond filenames

EmbeddingsFilesScoredRanked

Precision

Narrow by type, folder, and date

Only the files worth opening.

Combine the semantic query with filters for MIME type, folder, modification date, and trashed status so an agent retrieves exactly the documents that matter.

5
Filter Facets

Type, folder, date

MIME typeFolderDateTrashed

Governance

Private and on-premise

Your files stay your files.

Indexing and search run inside your perimeter, scoped per user with audit logging, so sensitive documents are safe to make searchable by an agent.

100%
On-Prem

Per-tenant, logged

On-premRBACAudit logSovereign
Inputs

Parameters

The google_drive_vector_search tool accepts these inputs when an agent calls it. Required inputs are flagged.

Name Type Required Description
query string Required Search query for semantic matching.
user_id integer Required User ID for multi-tenant isolation.
top_k integer
default: 10
Optional Maximum number of results to return (1–50).
mime_type string Optional Restrict results to files of this MIME type (e.g. application/pdf).
folder_id string Optional Restrict results to files within a specific folder.
date_from string Optional Only include files modified on or after this date (ISO 8601).
date_to string Optional Only include files modified on or before this date (ISO 8601).
is_trashed boolean Optional Include or exclude trashed files.
Where it pays back

Where Drive search pays back

Contract retrieval

Find the signed agreement by describing its terms, not its filename.

Proposal reuse

Surface a past proposal or deck to adapt instead of starting over.

Policy lookup

Find the current version of a policy document across shared drives.

Research recall

Locate the report that covered a topic without remembering where it was filed.

Onboarding

Let new hires find the right template or spec by describing what they need.

Agent grounding

Ground a knowledge or research agent in your real document store.

How VDF AI connects it

Assigned to agents, orchestrated as networks

On VDF AI, an industry’s use cases map to agents, and you assign tools like this one to those agents. Compose multiple agents into a governed, on-premise network.

ROI Snapshot

What changes after you assign it

Faster
Time to the right file
Filtered
By type, folder, and date
Ranked
By relevance, not name
100%
Searchable without files leaving
FAQ

Questions about the Google Drive Semantic Search tool

What is the Google Drive semantic search tool?

It searches your vectorized Google Drive files and folders by meaning and returns ranked matches with similarity scores. Assigned to an agent, it finds the right document even when filenames are meaningless.

Can I limit it to a folder or file type?

Yes. Provide folder_id to scope the search to a folder and mime_type to restrict to a file type such as PDFs — combined with the semantic query.

What filters are supported?

You can narrow by mime_type, folder_id, modification date range (date_from/date_to), and whether trashed files are included.

Are our files exposed?

No. Indexing and search run on-premise or in your sovereign cloud, scoped per user and audit-logged, so documents never leave your perimeter.

How does it pair with other tools?

It is often assigned alongside Gmail and calendar semantic search, and the file summarizer, so an agent can find a document and then summarize or act on it.

Let agents search your Drive by meaning

See Google Drive semantic search assigned to an agent that grounds answers in your real documents — on-premise.