In a world where the most absurd secrets are the true ones, the CIA's answer to bugging the Soviet embassy was not a gadget out of a spy novel. It was a cat. A real, breathing cat, surgically rebuilt into a four-legged listening device, in a Cold War project the Agency itself called Acoustic Kitty. And on its first outing, the multimillion-dollar feline operative was, by the most repeated account, killed by a taxi.

Here is what actually happened, stripped of the late-night-comedy gloss. In the 1960s, the CIA's Directorate of Science and Technology faced a hard problem: how do you plant a microphone where humans, dogs, and sweep teams can't go? Someone reasoned that a cat could wander up to a park bench, a wall, or an embassy garden and be ignored entirely. So the Agency had a veterinary surgeon implant a microphone in the cat's ear canal, a small radio transmitter at the base of its skull, and a fine antenna wire woven up its spine into its tail, with a battery to power the rig. The cat was, in effect, turned into a mobile, self-propelled bug.

The proof this happened is not folklore. It sits in a partially declassified CIA memorandum, 'Views on Trained Cats [redacted] for [redacted] Use,' which the National Security Archive at George Washington University obtained and published. The document is heavily redacted, but its surviving language is unmistakable: it praises the team's work, notes that the principles were sound, and concludes, with bureaucratic understatement, that the effort was 'a remarkable scientific achievement.' Former CIA officer Victor Marchetti later put the total program cost in the neighborhood of $20 million across roughly five years of development.

The famous ending, the part everyone repeats, is that on the cat's first real mission, near the Soviet compound on Wisconsin Avenue in Washington, the animal was released, wandered into the street, and was struck and killed by a taxi almost immediately. That detail comes largely from Marchetti's recollection rather than from the declassified paper itself, which is too redacted to confirm the taxi. So treat the taxi as plausible legend, not documented fact. What the document does make plain is that the program was eventually abandoned as impractical, because the fundamental flaw was never the engineering. It was the cat. You cannot order a cat where to go.

The skeptical-but-fair reading actually makes the story stranger, not tamer. Skeptics like to wave Acoustic Kitty away as a myth inflated by spy-buffs, but the core is genuinely confirmed by an official, FOIA-released CIA record. What's exaggerated is the periphery: the precise $20 million figure rests on one ex-officer's memory, and the cinematic taxi death is uncorroborated by the paperwork. The honest version is less a punchline and more a portrait of an intelligence culture so flush with black-budget money and Cold War paranoia that surgically converting a house cat into a transmitter passed internal review.

What lingers is what the redactions hide. The released memo blacks out exactly who the intended targets were, which foreign mission the cat was meant to penetrate, and what the Directorate of Science and Technology learned that it considered worth keeping classified decades later. Acoustic Kitty is funny right up until you ask why, after half a century, the Agency still won't say in full what it wanted its cat to hear.