← All integrations

Grok Bot and Google Drive

Google Drive 24 bots in the directory connect to Google Drive, and 1 guide covers setting it up. Each bot declares the one action it never takes without you, which for a connected tool is usually the write, the send or the payment.

Bots that connect to Google Drive

Ad Creative Generator
Turns one confirmed brief into six labelled ad variants as drafts, each carrying its angle, its tested hypothesis, and the source of every claim.
Boundary: Never launches a campaign, edits a live ad, or spends a credit; every variant is written to a draft file for a human to move.

Argue The Other Side
Takes a decision you have already made and builds the strongest case against it, then tells you what evidence would actually change your mind.
Boundary: Never makes the decision, never recommends a side, and never softens the counter-case because you appear committed.

Bookmark Skill Grader
Grades a month of bookmarks into keep, maybe, and noise, then turns keepers into spoken skill triggers you can paste.
Boundary: Never deletes bookmarks or installs skills; it grades and drafts, then waits for you to keep them.

Brand Deck Keeper
Audits every customer-facing deck against the live price book and changelog, then hands you paste-ready fixes for each stale slide it found.
Boundary: Never edits the master deck, its components, or the shared library; every fix is a proposal I apply myself.

Citation Checker
Opens every link and checks every quote in a draft before it goes out, reporting dead links, moved pages, and quotes that do not match the source.
Boundary: Never edits the draft and never silently swaps a dead link for a similar one. It reports and a human decides.

Claim Provenance Tracker
Traces every number in a deck or doc back to where it came from, and flags the ones whose only source is an older deck.
Boundary: Never edits a deck or a document, and never updates a number it believes is stale. It reports provenance and stops.

Demo Clip Library
Indexes every recorded demo by objection, feature and proof so you can find the ninety seconds that answers a prospect, each moment timestamped.
Boundary: Never sends a clip to a prospect, publishes a recording link outside the team, or deletes or trims a source recording.

Directory Page Drafter
Researches one niche listing a day into Markdown with cited sources, then stops so a human publishes the page.
Boundary: Never publishes, deploys, or pushes a live page; drafts stay in Markdown until you ship them.

Enablement Pack Builder
Assembles an internal enablement pack on any topic from your drive, call recordings and email threads, so a teammate can handle it tomorrow.
Boundary: Never sends the pack to a customer and never shares the file outside the workspace; it lands as an internal draft link for one teammate.

Expense Reconciler
Reconciles the week's expenses against receipts, flags policy exceptions with the rule they broke, and drafts the chase. It changes no reimbursement and sends no message.
Boundary: Never sends a message and never changes a reimbursement, an approval state, or an amount. It drafts and stops.

Expert Call Prep
Builds the question list for an expert call from what you already know and what previous calls left open, so you stop paying to re-ask answered questions.
Boundary: Never contacts the expert, never schedules the call, and never shares prior call notes outside the operator's own workspace.

First Draft From Outline
Turns an outline you already approved into a first draft that follows it exactly, and flags every place the outline was too thin to write from.
Boundary: Never publishes, never sends, and never invents a fact to fill a gap. Thin sections are flagged, not padded.

Grok Imagine Storyboard
Researches a story, plans scenes, and generates Grok Imagine shots into a private folder you review before anything goes public.
Boundary: Never posts, publishes, or shares unreviewed shots; the folder stays private until you release it.

Headline Variant Desk
Produces headline options for one finished piece, each labelled with the angle it takes and the promise it makes, so you pick deliberately rather than by feel.
Boundary: Never publishes or posts a headline, and never writes a promise the piece does not actually keep.

Jargon Stripper
Rewrites a draft for one named reader, listing every term it replaced and why, so you can put back the ones that were load bearing.
Boundary: Never publishes, and never removes a technical term without listing it, because some jargon is precision the reader needs.

Literature Scan
Surveys the published work on one written question, groups findings by what they actually agree on, and names the disagreements instead of averaging them away.
Boundary: Never presents a synthesis as consensus when the sources disagree, and never cites a paper it could not open.

Market Sizing Worksheet
Builds a sizing model where every assumption is a named, sourced line you can change, so the number is auditable rather than asserted.
Boundary: Never presents a single headline number without the assumption table beneath it, and never sources an assumption to itself.

Marketing OS Auditor
Scores a marketing system against a fixed rubric, drafts the next experiment, and never publishes, spends, or launches an ad.
Boundary: Never publishes, spends ad credit, or launches a campaign; audits and drafts wait for you.

Marketplace Listing Scout
Flags public marketplace listings that look underpriced versus a stated median, using photos not titles, and never messages the seller.
Boundary: Never messages a seller, never books pickup, and never buys; it files flags and waits.

Podcast Clip Desk
Cuts a timestamped YouTube or podcast span, pulls quotes, and drafts caption copy that waits for you to publish.
Boundary: Never publishes, schedules, or posts the clip; drafts stay in chat or Drive until you ship them.

Source Verifier
Takes a list of claims and checks each against a primary source, returning verified, contradicted, or unverifiable with the link and date. It never edits the document.
Boundary: Never edits the source document and never removes a claim. It returns a verdict list and a human decides what to change.

Style Guide Enforcer
Checks a draft against your written style guide and reports each departure with the rule it breaks and the line it happened on. It never rewrites the draft.
Boundary: Never rewrites, never silently corrects, and never applies a rule that is not written in the guide it was given.

What Did We Promise
Finds the commitments your team made to an account and nobody tracked, quotes each one word for word with a link, and flags the ones now overdue.
Boundary: Never contacts the account and never answers a promise on your behalf; the quoted list goes to the internal owner only.

What Should I Cook
Picks tonight's dinner from what is actually in your kitchen, ranked by what expires first, and never sends you shopping for one missing herb.
Boundary: Never orders groceries, never places a delivery order, and never adds anything to a cart. It suggests, you cook or you order.

Guides that cover Google Drive

Grok Bot and Google Drive: Permissions and What to Automate
Tutorial · A Grok Bot Google Drive setup where sharing is the danger, not editing: which access families to grant, how folder permissions spread, and a librarian charter.

Missing a Google Drive bot you would use? Add one: one markdown file, one pull request. Or search the whole catalogue for Google Drive.