I've found more success by taking a larger sampling of sources on seeing predictions, as they can definitely vary. In fact, I check 5 different seeing reports (plus 3 additional ones for clouds) to judge the night ahead of time. Occasionally, one or more of them can be very far off, which is why I use so many. However, usually the mean judgment between them all does very, very well. Here they are:
Meteoblue- This doesn't always give results that reflect how my imaging will be, however it contains one of the most important metrics anyways, which is the jet stream speed. Higher speeds (into the red), and it will definitely be worse. Lower speeds, better seeing. Being near mountains, the jet stream is almost always fast, and the seeing almost always terrible.
ClearDarkSky- One of the best, bookmark your location and I say its seeing estimates are good 80% of the time.
Apps:
"Clear Outside"- Great for weather, not heavily-weighted for my seeing predictions though.
"Astrospheric"- Pretty good seeing estimates as well. Discussed above.
"Astro Panel"- Okay predictions for seeing.
Viewing the cleardarksky and astrospheric predictions, coupled with the jet stream information, I can usually know what to expect. I'll say though, it's not an easy science; some of my best seeing was during mediocre forecasts. If you're feeling it, try anyways!
I also think looking at the stars does an excellent job of predicting how well the guider would do from seeing. Every single night I do multiple seeing assessments by rating the sky between 1 and 5 mentally, by looking at it, and I think I could reliably predict my RMS error values on PHD2 just by looking at the stars (only look in the area you're imaging; it can vary widely throughout the sky!).
Also, always image as high in the sky as is possible. Always.