Instrument identifier. Upload audio and AI lists every instrument it hears, including layered and processed sounds.
Select the AI model for audio analysis. Different models may have different capabilities.
Record audio directly from your microphone
To identify the instruments in a song, upload the audio and the AI lists every instrument and sound source it hears, with the specificity that actually matters to producers: Rhodes versus grand piano, 808 versus acoustic kick, not just piano and drums. For each instrument it estimates prominence in the mix (lead, supporting, or background), describes any processing and effects it detects (distortion, reverb, chorus, pitch shifting, sample manipulation), and judges whether the source is acoustic, electric, synthesized, or sampled. It also flags unusual or signature sounds, the textures that define a track but resist naming. This solves the constant producer problem of hearing a sound you want and having no vocabulary to search for it. Layered and heavily processed sounds are exactly where human ears struggle, since a detuned synth stacked with a sampled guitar reads as one hybrid texture, and the analysis pulls those layers apart.
Upload the track and the AI dissects the arrangement layer by layer. Source identification names each instrument or sound as precisely as the audio allows, distinguishing closely related sounds like electric piano types, kick drum families, or string sections versus string pads. Prominence estimation places each element in the mix hierarchy: lead, supporting, or background. Processing analysis listens for the effects shaping each sound, including distortion, reverb, chorus, pitch shifting, and sample manipulation like chopping or time-stretching. Origin classification judges whether each element is acoustic, electric, synthesized, or sampled, which tells you whether to reach for a mic, a plugin, or a sample pack to recreate it. Signature sound flagging highlights anything unusual that gives the track its identity. The result is essentially a reverse-engineered instrumentation sheet for the arrangement.
Upload the track and the AI inventories every sound source it can hear, placed in the mix hierarchy (lead, supporting, background) with notes on processing and whether each source is acoustic, electric, synthesized, or sampled. Signature and unusual sounds get flagged separately, since those are usually the ones that made you ask.
Producer-level specific where the audio allows: Rhodes versus Wurlitzer versus grand piano, 808 versus acoustic kick, strings versus string pad, fretless versus upright bass. Generic answers like keyboard and drums are useless for recreating a sound, so the analysis commits to the closest specific identification and says when it is uncertain.
Yes, electronic sources are half the job. It distinguishes synthesized from sampled from acoustic, describes synth character (analog-style warmth, FM glassiness, supersaw stacks), and notes sample manipulation like pitch shifting or time stretching. You will not get the exact preset name, but you get enough description to rebuild the sound.
Background layers are specifically part of the analysis, and it routinely surfaces pads, ghost percussion, and doubling that casual listening misses. There are physical limits: sounds mixed near-identical to each other or buried under heavy masking may be described by their combined texture rather than separated into individual instruments.
Reliable on exposed, characteristic sounds; appropriately hedged on dense or heavily processed material. Distortion, extreme effects chains, and layering blur the line between sources, and a hybrid sound (a kick layered from three samples) resists a single name. The analysis says what it is confident about and offers best guesses for the rest.
Mostly recreation and learning: producers reverse-engineering a reference track, musicians figuring out what to cover a song with, sample hunters confirming what they are hearing, and curious listeners settling arguments about whether that is a real horn section. The prominence and processing notes turn the list into a usable arrangement map.
Do we have rapport? Upload a conversation for AI scoring of mirroring, timing, and style matching.
How high is my EQ? Upload a recording for AI to evaluate empathy and emotional awareness in your voice.
Is this audio AI-generated? Upload a clip and AI scans for synthetic artifacts and deepfake indicators.
What is my cat saying? Upload meows, purrs, or chirps and AI decodes feline emotions and needs.
What is my dog saying? Upload barks, whines, or growls and AI decodes canine emotions and needs.
What bird is this? Upload a call or song and AI identifies the species and decodes the meaning.