Privacy & data

Privacy & Data

Data transparency

All snapshot files are available for download from the data page. Every published ranking archives the exact board data it was computed from — the full snapshot history is downloadable.

The methodology is fully documented and the formula is open. Anyone can replicate our scores from the same snapshots.

No data is collected from visitors. The only data in this project is the board data fetched from public leaderboards.

No tracking

This site uses no analytics scripts, no cookies, no tracking pixels, no fingerprinting, no embedded third-party content that phones home.

The only third-party resource is @vercel/speed-insights (Vercel's optional speed monitoring, not analytics). Even that is minimal and privacy-respecting.

This is an intentional choice. The purpose of this board is to eliminate bias in TTS discovery, not to surveil its audience.

Purpose

The goal is simple: aggregate real quality data so consumers can make cost/value decisions based on unbiased information.

We are here to eliminate the bias that established models get from paid content and bought "neutral" coverage in major outlets. If a model is top on all three independent boards, that's signal.

We also aim to benefit new researchers and smaller vendors by reducing the time to discovery. A model that's strong on one board but not yet on the others shows up in our watchlist immediately — no media campaign required.

Project status

This is a hobby project. The domain is sponsored by a benefactor. There is no commercial interest, no vendor affiliation, no pay-to-play.

Privacy is respected the same way it is in other leaderboard and aggregator projects — maybe more so, given the explicit no-tracking stance.

Contact: texttospeechdotcom@gmail.com