How this works

draglessdrift turns today's real water conditions into a specific trout fly, technique, and gear recommendation — with the reasoning shown, not hidden. It's built to be another data point in your planning, not a replacement for a guide, a local shop, or your own judgment.

An angler wading a wooded Pennsylvania creek, net in hand, fishing the Sinnemahoning.

What the recommendation is grounded in

Water flow and temperature, from USGS. Real-time streamflow, discharge, and water temperature from the USGS Water Services API — the same gauge data anglers already check by hand. This is what actually shapes the recommendation.

Pressure and precipitation, from NOAA. Live barometric pressure and recent precipitation from the National Weather Service. Shown for context on the results page — they don't change the recommendation itself.

Hatch charts and technique notes, curated by hand. A written file per water — hatch timing, named fly patterns, and technique guidance we've researched and compiled ourselves, not scraped or auto-generated. This is what gets combined with the live conditions above to produce the recommendation.

Where this works today

draglessdrift currently covers 176 waters (183 sections) across Pennsylvania, New York, and Virginia. We're expanding from there.

How a recommendation gets made

1. We pull today's live flow and water temperature from the USGS gauge tied to your section, and check how those numbers compare to that gauge's own decades of recorded history.
2. If water temperature is 68°F or higher — the threshold where catch-and-release starts to risk becoming catch-and-kill — we stop there and don't generate a fly recommendation. Instead we point you to a sibling section on the same water that's confirmed cooler. If no temperature reading is available at all during peak summer risk months, we say so plainly instead of assuming it's safe.
3. Otherwise, that water's knowledge-base file — the hatch chart and technique notes described above — gets handed to Claude, the AI model behind this app, along with today's date and the live conditions. Claude reads the file directly, the same way a person would, and reasons about which hatches are realistically active right now, given the date and the numbers.
4. That reasoning produces a ranked list of up to five flies, each with a technique, rod weight, line, leader, and tippet spec, plus a written rationale you can judge for yourself. It's never padded to five if the water's data doesn't genuinely support that many, and it's never allowed to invent a flow or temperature figure it wasn't given.

Data & privacy

No accounts, no tracking of who you are. We use Vercel Analytics for aggregate, cookieless traffic counts — nothing about your searches is stored or sold.

Something wrong, missing, or worth flagging? Feedback

draglessdrift is one data point — a specific, honest starting point built from real conditions. It's not a replacement for a local guide, a shop's advice, or your own read of the water.

Ready to see it for your water?

Search a water