Verify Watermark

The public AI detection tool. Ask whether a clip carries the watermark Speechify seals into audio it generates, with no account, no API key and no credential of any kind. `verify` answers; `detect` measures. This route returns a bare yes or no, the way verifying a signature does. Its sibling `POST /v1/audio/watermark/detect` takes an API key and returns the detector's confidence alongside the verdict. This is the programmatic half of the tool published at <https://speechify.ai/detect>, and it exists so the tool can be invoked without visiting our website, as California's AI Transparency Act (BPC 22757.2) requires. Nothing about the clip is stored, and nothing identifying about you is collected or retained. The answer is a bare verdict. `watermarked: true` is positive evidence that the audio came from Speechify synthesis. `watermarked: false` is NOT proof that it did not: only models redeployed since the watermark shipped mark their output, the detector needs at least three seconds of clear speech to judge, and re-encoding or changing the speed of a clip degrades the mark. Treat a negative as the absence of evidence rather than as evidence of absence. Because the tool takes no credential, it is rate-limited per client address and shares a platform-wide budget: expect a 429 under sustained automated use, and retry after the interval the response advertises. Use `POST /v1/audio/watermark/detect` with an API key for the detector's confidence score and a per-workspace allowance of its own.