Walk the floor. Speak the punch.
The Punch List Drafter is the Trueleveler engine that turns a site walk's voice notes and photos into a structured punch list. Voice + photo capture during the walkthrough. Trueleveler transcribes, clusters items by room, tags trade and severity, attaches the photo with location metadata, and matches each item to the responsible subcontractor from your vendor directory with a due date. Nothing is sent to the sub until you send it. The punch list writes itself between the walk and the next coffee break.
Try it now
Punch List Drafter runs on a site-walk recording and needs a free account (no card). It has no sample data yet, and the no-signup demo only levels bids.
Walk · talk · photo.
Three steps on the mobile app. Total field time on a 200-unit residential floor: ~30 minutes. The desk work that used to take a full afternoon — transcribing notes, organizing by room, assigning to subs, building the punch list — happens automatically while you walk to the parking lot. Export it as PDF or CSV and send it whenever you're ready. There's a second capture path, too: drop a photo pin on the drawing and it becomes a punch item at that exact spot — same trade tagging, same sub matching.
Photo + voice tag
Take a photo. Hold the mic. "Unit 4B kitchen — cabinet door misaligned upper left corner, also baseboard caulk gap at the dishwasher." Geotag + EXIF lock the location. Two items captured in 6 seconds.
Auto-group by room
The engine checks each item's location against every room already logged on this project and reuses the exact same label when it matches — so three items in "Unit 4B Kitchen" show up as the same room, not three near-identical phrasings. A photo's GPS tag helps confirm a match when the phone captured it (indoors, many won't), but the transcript always wins. That consistency is what lets the punch list actually sort and group by room, floor after floor.
Trade · priority · assignee
Each item gets tagged with trade (cabinetry, paint, drywall, HVAC), priority (low / medium / high / critical — critical reserved for safety or occupancy-blocking items), and the responsible sub matched from the project's vendor directory. It lands on your punch list as an open item, assigned to you, with a due date (default: 14+ days for low, 7 days for medium, 1–3 days for high/critical). Nothing goes to the sub until you export the list and send it.
From a 30-minute walk. To a punch list ready to send.
One captured walk produces a structured punch list per floor, sorted by location, with photo evidence, trade tags and a suggested sub already on each item.
A punch list grouped by sub, by floor, by priority — with photo evidence and deadline per item. Export it as PDF or CSV and send it to each sub however you normally do. The PM gets a project-wide dashboard showing punch density by floor, by trade, and by sub-completion-rate.
The walk used to be step one of three.
Before this engine, the punch-list workflow was three jobs that one person had to do:
~3-4 hours / floor
Walk every unit, every common area. Scribble on a clipboard or type into a tablet. Phone photos in the camera roll, hopefully captioned. Mental note: which trade owns each item.
~4-6 hours at the desk
Transcribe handwritten notes. Match photos to items. Type into a spreadsheet. Group by sub. Set deadlines. Forget half of what you saw because you didn't write it down clearly.
~1-2 hours
Email each sub their list. Copy the PM. Set follow-ups. Track completion in a separate system. Send reminders before the walk-back.
30 minutes · the walk itself
Job 2 disappears and Job 3 shrinks to one step. The engine clusters the items, matches subs, and sets due dates while you finish walking, so the list is ready to export and send. The PM reviews it when they get back to the trailer — not the next afternoon.
Times above are typical for a residential mid-rise (200-300 units). Smaller jobs scale down proportionally; healthcare and lab work scales up because of code-driven items per room.
The list closes itself.
Punch items live as Tasks in the project's tracker. Each sub gets a per-sub view (their items only). Each item gets:
- Verify on completion — sub takes a photo of the corrected work, uploads it inline on the item. PM gets notified and closes it themselves after review — automatic before/after photo-match closing is on the roadmap, not in the product today.
- Walk-back ready — when 80%+ of items are closed, the engine schedules a follow-up walk and pre-loads the open items for re-verification.
- Sub scorecard — per-sub completion rate, average time-to-close, and quality (re-open rate) feed the Vendor Profile aggregator for future bid scoring.
It works alone. It's better in a chain.
Punch List is a closeout-phase engine. Its output feeds two downstream chains: the closeout package and the sub-vendor scorecard.
- Meetings — pre-CO walkthroughs recorded as meetings auto-extract punch items into the list.
- Closeout Builder — all-zero punch list is a milestone the closeout package needs before owner handoff.
- Tasks — punch items live as tasks; per-sub completion rates roll up into the project dashboard.
- Vendor profiles — close rates + re-open rates per sub feed their reliability score for future bidding.
Things everyone asks first.
What about offline / spotty cell in the building?
Capture works fully offline. Photos and text save locally and sync automatically the moment connectivity returns. Voice notes save locally too; if an upload fails, a held-recording prompt lets you retry it with one tap once you're back online. Most walkthroughs in concrete/steel structures lose cell deep inside — this is the expected case, not the edge case.
How accurate is the auto-trade tagging?
The engine tags trade from what the PM actually said in the transcript and photo captions — if the walk-through doesn't make the trade clear, it leaves the field blank rather than guess. The PM can set or override trade assignment with one tap on any item. Low-confidence items get flagged for PM review before they ever reach the exported list.
Can owner-rep walk it too, or just GC?
Yes. The engine works the same for whoever runs it: the drafted items land on the punch list of the person who ran the walk, and nothing is sent to anyone automatically. So an owner-rep can export the list and hand it to the GC PM, who reviews it and sends it to the subs. Useful for owner-side reps who want their walk to be authoritative without bypassing the GC's command structure.
Does it integrate with Procore / Autodesk?
Export-only for v1. Exports punch lists as PDF + structured CSV that you can drop into Procore's Inspections module or Autodesk Build's Issues. Direct API integration with both is on the roadmap.
Walk a floor with us. The list closes itself.
Bring an actual walkthrough to a 15-minute call. We'll capture it live on screen and generate the punch list. You keep the output.
Try Free for 14 Days →