Use cases

Who Sery is built for

Three lives where the documents that matter — contracts, charts, workpapers, family records — can't leave the office, the laptop, or the partner's machine. Pick the one that sounds like you.

~/code · zsh
~ $ sery search customer_id
3 sources · 4 matches · 280 ms
s3://analytics/users.parquet
column: customer_id · BIGINT · 4.2M rows
sftp://hetzner-archive/dump.csv
column: customer_id · VARCHAR · 12,481 rows
drive://Receipts/clients.xlsx
column: customer_id · STRING · Sheet 1 · 234 rows
b2://backups/q3-2025.parquet
column: cust_id · BIGINT · 1.1M rows
Files never left their storage. SQL ran where the bytes already lived.
The developer

Read every file across S3, SFTP, Drive, and B2 in one search

Free forever (unlimited sources, fully offline) · $19.99/mo for AI across every file

Scenario

A backend engineer at a Series B SaaS. Daily-driver MacBook. Company S3 for analytics dumps. Personal Backblaze B2 for off-site backup. A side-project SFTP archive on a Hetzner box. A Google Drive folder of vendor PDFs, security reports, and contracts they can't paste into a cloud LLM. Four tools, every day.

Challenge

The AWS Console is slow and per-bucket. Cloud browsers show file lists but can't read a Parquet without downloading 1.4GB. Drive web doesn't search inside PDFs. No single tool reads tabular data AND vendor PDFs AND searches across all of them, let alone reasons about them locally.

Solution

Install Sery Link — your private data agent on the desktop. Connect every storage in five minutes, credentials in the OS keychain. PDFs and contracts read in-process via pdfium (OS-native OCR for scans). Tabular files preview in 2s via the Parquet footer. Search 'customer_id' or 'indemnification' → hits across S3 + SFTP + Drive simultaneously.

Result

Four tabs replaced with one app. Read PDFs and Parquets on your machine. Search content + columns across every source. Vendor contracts never go to a cloud LLM; tabular bytes never leave S3.

app.sery.ai · Network
You — Sunday 21:14

find my 2024 accountant letter — should be a PDF or DOCX

MacBook · work
8,402
Desktop · home
12,109
Synology · NAS
847,233
workspace · 867,744 files
Sery — found 1 match across 3 machines

accountant_2024.pdf

On Synology NAS · /backups/Tax/2024 · last opened Jan 8

The power user

Personal documents across three machines, one private workspace

Free covers each machine's agent · $19.99/mo unlocks AI + cross-machine

Scenario

A senior designer. MacBook for work, Windows desktop at home, Synology NAS for photos and tax archives. Mortgage paperwork, accountant letters, school records for the kids, every receipt PDF since 2019. Three machines, one life.

Challenge

OS-level file search only sees one machine — and only filenames, never the PDF contents. Sync-style cloud storage wants to upload everything. They long ago abandoned tools that put tax returns on someone else's server.

Solution

Install Sery Link on each machine and pair them with one workspace key. PDFs and DOCX read in-process on the device that owns them. Ask from any laptop at app.sery.ai: 'find the 2024 accountant letter', 'summarise the mortgage refi paperwork', or 'draft a school-transfer letter using my kid's most recent report card.' The answer assembles across all three machines.

Result

Documents never move. The machines they already owned finally act like one private workspace — find, summarise, and draft across all of them, without any file leaving the laptop or NAS it lives on.

Workspace admin · Mason & Co · CPA
Workspace status
All client data is on the partners' machines. Nothing in any cloud.
0 KB uploaded
Connected machines
Sarah Mason · MacBook
12,802 files
James Wu · ThinkPad
8,447 files
Priya Patel · iMac
4,109 files
Office · file server
44,920 files
Recent activity
14:22Jamesqueriedclients_q3.csv
14:08Sarahran recipe"missing-q3-data"
13:51Priyaindexedacme-receipts/
13:40Officeindexed2024-tax-returns/
The small firm owner

A 3-person practice puts AI on client files — without uploading any

Plus today (5 invited machines on one bill) · Team tier when validated

Scenario

A partner at a small CPA firm. Two juniors work from home and the office interchangeably. Clients email PDFs of 1099s, prior-year returns, K-1s, and brokerage statements. Tax season means quarterly panic across three laptops plus the office file server, with documents in every email thread.

Challenge

Dropbox Business sees every file but can't read a 1099 PDF, summarise a brokerage statement, or draft a planning memo. Building a shared database is a year of work they won't do. Clients will absolutely not tolerate their tax records going anywhere near a third-party cloud AI.

Solution

Install Sery on every partner's machine and the office server. AI reads the client PDFs in-process on whichever machine they live on. Workspace recipes for common client work ('categorise this client's transactions', 'draft a year-end planning memo'). Partner dashboard shows which machines are connected and who queried what, when.

Result

AI-drafted summaries, memos, and answers grounded in client files — without a single PDF leaving anyone's laptop. Audit log shows exactly who read what, when.

Where this goes next

Vertical SaaS companies that embed Sery instead of building their own AI layer

Once Sery is deployed across enough verticals, the SaaS vendors those customers use — specialty CRMs, discovery platforms, compliance tools — start asking: can we embed Sery in our product so our customers get the AI layer without us having to build 18 months of infrastructure? That's the platform tier.

Embed SDK + public API arrive in 2029. Partner program details on the contact page for early-interest conversations.

Start where your documents are

Free forever — read PDFs and tabular files across unlimited sources and unlimited machines, fully offline. $19.99/month adds AI that finds, summarises, drafts, compares, and reasons across them — 7-day trial, no card.