Temperature is the Ghost Kitchen & Dark Kitchen delivery complaint you can actually engineer out: hold in range, pack to stay in range, and probe at the moment of sealing. These 8 checks cover four things. Log chilled storage and every brand's prep rail at 0–5°C (32–41°F). Hold hot food at 63°C (145°F) or above. Probe one hot item per time of day at seal. Separate remakes from ordinary waste. The evidence includes an MDPI trial of 249 hot ready-to-eat meals under simulated delivery. Internal temperature fell to 40°C (104°F) in 45 minutes on a single-layer tray. It took 140 minutes stacked six deep.
In a trial of 249 portions of hot ready-to-eat meals under simulated delivery at external temperatures of −2°C to +33°C (28–91°F), internal temperature fell to 40°C (104°F) in 60 to 140 minutes, depending on stacking. This took about 45 minutes for a single-layer tray, about 100 minutes at three layers, and about 140 minutes at six layers. Polypropylene versus sugarcane bagasse trays differed by 50 minutes. Bacillus cereus averaged 2.99 log CFU/g after four hours' culturing at 40°C (104°F), from undetectable initial counts. The authors concluded that hot meals delivered within four hours under these conditions are unlikely to let bacteria grow to unsafe levels. Packaging material and stack depth are the biggest things the kitchen can control.
Chilled storage and every brand's prep rail run at 0–5°C (32–41°F). Anything held above 8°C (46°F) is neither packed nor sold. Hot-holding runs at 63°C (145°F) or above, or 57°C (135°F) where your local code sets that line. Multi-brand sites need each rail logged separately. More brands means more rails and more places to drift. Meat held at unsafe temperatures was among the cited violations across 61 inspected multi-brand ghost-kitchen units, 58 of which had violated rules. Hot ready-to-eat food is on a four-hour clock from the moment it leaves temperature control. In a delivery-only kitchen, that clock includes time sitting sealed on the handoff shelf.
You cannot, and no one publishes that number. The available temperature evidence is either simulated delivery or meal-kit cold-chain data on keeping food cold during delivery. No dataset measures hot-food temperature at the door across real third-party orders. The practical substitute is to probe one hot item per time of day at the moment of sealing, and log the reading. Then compare it with your cold-food complaints. Independent audits give you the outcome side of the equation. Temperature accuracy ran 91% for direct deliveries versus 83% for batched ones. Satisfied customers received their orders 8 min 54 s faster than unsatisfied ones.
One free app has all of them: open the Ghost Kitchen & Dark Kitchen Checklists app - offline, no account, add it to your home screen.
Every statistic on this page comes from public inspection and food-safety data: