Resumable Scraper Kit
terminal
$ python my_scraper.py
got page 3739 [checkpoint saved]
got page 3740 [checkpoint saved]
^C — killed mid-run
$ python my_scraper.py
resuming: 3 741 done, continuing at 3 742
…
second run: 0 new items — delta working
Every scraper dies eventually — a 429, a dropped connection, a layout change.
The difference is what happens next: this kit resumes from a checkpoint instead of starting over,
backs off on rate limits, and on repeat runs returns only what's new.
- Python skeleton — requests-based, no Selenium
- Checkpoint & retry engine, snapshot-delta pattern
- Two worked examples + 1-hour adaptation guide
Came out of job-board and marketplace monitors that run for me every morning.
Doc-Pipeline: inbox → document
email in → docx out
arrives"Hi, please find our order —
ACME Supply order #A-1042,
total 18 450.00 USD,
net 30, deliver to Dayton…"
→
filed, 40 sec laterorder_A-1042.docx
Company: ACME Supply
Amount: 18 450.00 USD
Terms: Net 30
→ Drive/Incoming orders/
Orders and invoices arrive as email; someone retypes them by hand.
This n8n workflow watches the mailbox, extracts the fields with Claude and produces
a formatted docx from your own template. Unrecognized mail lands on an exceptions
sheet — nothing is lost silently.
- Full n8n workflow — import and run
- docx from your Google-Docs template + Drive filing
- Idempotent: runs twice, nobody gets duplicated
Born in a logistics back office that retyped supplier orders daily.
Telegram Site-Report Bot
what the crew sees
poured concrete 3rd floor section B, 12 CY, me and Jake, 9 hrs 📷
📋 Riverside School
🔨 Concrete — poured 3rd floor, section B
📏 Qty: 12 CY · 👷 2 crew · 9 h
📎 photo filed
Your crew already texts you what got done — the bot turns those messages
into a real site record. A dated daily-log entry out of every text: job, work item,
quantity, photo evidence. When a pay application gets questioned or an extra gets disputed
months later, you have a timestamped paper trail — not a scroll of chat.
- No forms, no app for the crew to learn — free text + photos
- Daily log / site diary builds itself: Excel summaries + dated photo index
- Python source (aiogram), pluggable trade vocabulary, your own VPS
This kit came out of a reporting system used daily on a 1,100-student school build — 4 crews, every day.
English kit is being finalized — email me and you'll get it first, at the launch price.