Files each new upload into the right folder from its name, type and metadata, comments on the file explaining what matched, and leaves anything it cannot place exactly where it is.
Run a staged conversation — no account needed. The agent handles it for real while a simulated world answers its tool calls; nothing touches real accounts, and nothing is actually sent.
INV-2026-0417 Halvorsen Marine.pdf
Added by Ibrahim Diallo
PDF, 214 KB, single page, uploaded at 09:12 from the office scanner app. Created and modified timestamps match. Owned by the uploader and shared with the finance group. Eleven other files in the tree already begin with INV-2026, and a client folder named Halvorsen Marine AS sits two levels down under Clients.
Using this template drops you into a guided setup. It asks exactly this, nothing else:
Connect Google Drive
One sign-in. The agent acts through your account, scoped to what this template uses.
Connect Slack
One sign-in. The agent acts through your account, scoped to what this template uses.
Filing rules
Which folder is the intake, how you want files sorted from their names and types, the exact destination folder for each kind of file, and which uploads are worth a Slack ping.
Choose which agent runs it
NoClick's built-in models work out of the box — or bring Claude Code, Codex, and other coding agents on your own subscription.
Watch it handle a test run
A staged conversation against a simulated world — then it’s live.
A shared Drive folder becomes a dumping ground within a month, and the cost of that lands on whoever needs the invoice at quarter end. This agent picks up each new file, reads its name, type and metadata, moves it into the folder your rules name, and records on the file itself which rule put it there. It works from metadata alone because it genuinely cannot open your documents, so anything it cannot place with confidence stays exactly where it is with a question attached.
Can it read what is inside my files?
No, and it says so rather than pretending otherwise. It sees the file name, type, size, owner, timestamps and the folder tree around it, which is enough to sort by naming pattern, client code or file type, and not enough to sort by what a document actually contains. If your uploads are all called Scan 2026-08-11.pdf, fix the naming first or it will keep asking who they belong to.
What happens when it files something in the wrong place?
A move is reversible and the comment on the file records where it came from, so putting it back takes seconds. The genuinely damaging case, a confident guess on ambiguous metadata, is the one it refuses: unclear files are left untouched with a question. Point it at a copied folder for a week if you want to watch it before it touches the real tree.
How much does it run on a busy folder?
One short run per file change, so a team dropping forty files a week is forty runs, and a scanner that saves three hundred pages as separate PDFs is three hundred. Nothing runs while the folder is quiet. If the volume is heavy, watch the intake folder only rather than the whole drive, since every change in the watched tree counts.
Free to start. Guided setup, a test run against staged conversations, and it's live.