Gandr

Gandr is building an unmetered text-to-speech API: a flat-rate monthly subscription per 'stream' that lets developers generate completely unlimited speech, rather than paying per character or token. Its proprietary Spex-TTS engine covers 23 languages, offers instant voice cloning from roughly ten seconds of audio, and streams with low latency for real-time use.

As AI voice agents, receptionists and dispatch systems start speaking for hours a day, per-token pricing turns continuous audio into an unpredictable, escalating bill. Gandr's flat-rate model is a direct bet that high-volume voice applications need cost certainty, not a running meter.

Founded in 2026 and based in San Francisco, Gandr is being built around Hameed A. Al-Ghaithi and a founding team now hiring its first engineering and growth roles. Beyond that, the company shares little about its founders' backgrounds, so the pedigree is best treated as an open question.

With hiring still pointing applications directly to the founders, this is a lean, founder-led effort in its earliest days. The team's credibility will rest on Spex-TTS itself — the 23-language coverage, ten-second cloning and low-latency streaming it says it has shipped.