How long does a building permit actually take?
Everyone asks. Almost nobody measures it. Below is the median time from application to issue, by trade, computed on 166,391 real permits issued in Austin, Texas over the last three years.
The headline: turnaround is not one number. Among the six largest samples here, mechanical (hvac) clears in 7 days (n=18,429) while new construction takes 140 days (n=12,265) — about a 20x spread on well-populated trades. Across every category measured the range widens to 2–140 days (70x), but the thin samples at each end deserve less weight than the middle of the table.
Median permit turnaround by trade — Austin, TX
| Trade / category | Permits measured | Median days | 90th percentile |
|---|---|---|---|
| Foundation | 2,356 | 2 | 20 |
| Renovation | 3,216 | 6 | 210 |
| HVAC | 3,986 | 7 | 161 |
| Mechanical (HVAC) | 18,429 | 7 | 205 |
| EV charger | 1,124 | 7 | 24 |
| Solar | 3,646 | 8 | 76 |
| Plumbing | 45,059 | 17 | 351 |
| Electrical | 46,042 | 19 | 302 |
| Demolition | 3,007 | 22 | 242 |
| Roofing | 2,716 | 33 | 560 |
| Other / uncategorized | 11,278 | 64 | 582 |
| Pool | 2,349 | 73 | 507 |
| Interior remodel | 7,753 | 77 | 328 |
| New construction | 12,265 | 140 | 716 |
| Addition | 3,165 | 140 | 509 |
Measured 2026-08-05. Median = 50th percentile of (date issued − date filed), permits issued in the trailing three years, categories with at least 300 measurable permits. The 90th-percentile column is the one worth reading before you promise a customer a date: half of addition permits clear in 140 days, but one in ten takes 509 days or more.
Why the 90th percentile matters more than the median
A median tells you the typical case. Scheduling risk lives in the tail. Several trades here have a median under three weeks and a 90th percentile past six months — the same permit type can be routine or a quarter-long wait depending on plan review, corrections and inspection backlog. If you are staffing crews or quoting lead times, the tail is the number that burns you.
How this was computed
Turnaround is date_issued − date_filed. Only permits where the source publishes
both dates are counted, which is why the per-trade sample sizes differ. Permits where the
issue date precedes the filing date are excluded as source errors rather than silently clamped.
Austin is used here because it publishes both dates reliably across every trade; not every
jurisdiction does, and PermitStack labels which ones do.
Run it on your own city
Every field above is in the public API. Pull filed and issued dates for any covered city and compute the same table:
curl -H "X-API-Key: pk_your_key" \
"https://api.permit-stack.com/v1/permits/search?city=Austin&state=TX&category=hvac&per_page=100"
See the API reference for the full parameter list, or coverage for which jurisdictions publish both dates.