External seller information. PRN does not sell this product and does not imply endorsement.
ITC word-generation device mapping environmental readings to a phonetic and word database output
External seller
Seller & availability
External seller information. PRN does not sell this product and does not imply endorsement.
This product is sold by an external seller. PRN hosts directory information only and does not process purchases.
Field reference
Potential Field-Use Benefits
A device that converts sensor readings into spoken words drawn from a built-in dictionary. Because output is selected from a preset word list, any apparent relevance is easily explained by chance and should be treated as such.
PRN has not tested or reviewed this product. This information is provided for reference only.
Two different mechanisms sit in this category:
Spirit boxes / ghost boxes (HexBox, P-SB7, P-SB11) are radios modified to sweep rapidly and continuously through the AM and/or FM band instead of locking onto one station. As the tuner sweeps, it throws out a fast stream of broken audio fragments and static — slivers of real broadcasts, signal bleed and noise — which the listener hears as choppy, half-formed sound.
Ovilus-type devices do not transmit voices at all. They read environmental sensors (such as EMF and temperature) and map sensor states to entries in a stored word bank, then speak or display the selected word. The vocabulary is pre-programmed by the manufacturer; the device is choosing words from a fixed list based on sensor input.
In both cases the device is producing output from ordinary inputs — radio noise, or sensor-to-word lookup. PRN treats the output as raw material to be examined, not as a message.
Use with care
Limitations
Spirit boxes are, by design, receiving radio. A swept tuner picks up real stations and, on rapid sweeps, adjacent stations bleed together so a single perceived syllable can be a blend of two broadcasters — easily mistaken for a context-fitting word.
Auditory pareidolia does the rest. Broken fragments plus an expectation ("what's your name?") push the brain to resolve noise into name-like or answer-like words. This is normal perception, not evidence of a source.
There is no standard way to separate radio bleed from a "response," which is the core methodological problem with the device class.
Ovilus output is bounded by its word bank and its sensor triggers. It can only ever say words the manufacturer loaded, and a "relevant" word is selected by sensor state plus chance, not by comprehension.
Read the data critically
Common false positives
Real radio stations and station bleed captured during the sweep.
Pareidolia — hearing a word in noise, strongly amplified when the listener knows the question.
Priming / expectation — telling listeners what to listen for makes them "hear" it.
Ambient sound leaking into the box's small speaker or the room recording.
Ovilus: coincidental word hits from a fixed vocabulary; sensor noise driving word selection.
Best practice: run a baseline sweep in the same spot before the session, note nearby transmitters, compare across different sweep ranges, never feed the answer to the listeners, and corroborate with independent data before treating any word as meaningful. Skeptical analyses (including of the original "Frank's Box") attribute spirit-box output to ordinary radio reception and pareidolia; no controlled evidence shows these devices carry communication.
PRN has not tested or reviewed this product. This guidance describes the device class and is provided for reference only.
Paranormal Response Network is not a seller, reseller, certifier, or safety authority for any equipment shown here. Listings may include vendor-submitted, sponsored, affiliate-linked, imported, or externally sourced information. Presence in this directory does not mean PRN has tested, endorsed, or approved any product or vendor.