Gandr
Verified Oct 15, 2026 ActiveText-to-speech and voice-cloning API billed by concurrent stream instead of by character or minute.
Listed on EffectHub ·
Pricing
- Pricing model
- Freemium
- Paid options from
- $10/mo
- Billing
- Monthly
Free tier starts with 50,000 tokens (one token is one character). Take a Gandr is $10/month for 1M tokens, Pro $50/month for 5M; the unmetered Scale stream is $150/month billed annually ($1,800) or $180 month-to-month, extra streams $150 each, burst $10 per stream-day; Enterprise from 500 lines is quoted.
Use tool ↗About Gandr
Gandr is a developer-facing text-to-speech and voice-cloning API for live uses such as phone agents, narration and dubbing rather than one-off audio exports. It speaks 23 languages, clones a voice from roughly ten seconds of reference audio with no training job or per-voice fee, and quotes 116 ms server-side and 146 ms over-the-internet time to first audio. Agents connect over WebSocket, batch jobs over HTTP, and it plugs into LiveKit, Pipecat and Vapi. Small plans meter tokens, where one token is one character; the Scale plan instead sells an unmetered concurrent stream at a flat monthly rate, with extra streams and burst capacity priced per stream. It states SOC 2 Type 2, HIPAA-verified and GDPR-ready infrastructure, says nothing sent is used for training, and watermarks output inaudibly. It is backed by Y Combinator, but no legal entity or address appears on the site.
- Platforms
- Web · API
- API
- Yes
- Last verified
- Oct 15, 2026
Verified sources
- Pricing page: Free 50,000 tokens; Take a Gandr $10/mo for 1M tokens; Pro $50/mo for 5M; Scale one unmetered stream at $150/mo on annual ($1,800 billed yearly) or $180 month to month; volume streams $150 each, burst $10 per stream-day; Enterprise from 500 lines (15 October 2026)gandr.ai ↗
- Homepage: TTS in 23 languages, ten-second voice cloning with no training job, 116 ms p50 server-side and 146 ms internet time to first audio, LiveKit, Pipecat and Vapi integrations, SOC 2 Type 2 / HIPAA / GDPR infrastructure claims, Y Combinator backing; footer has no legal entity or address (15 October 2026)gandr.ai ↗
- Terms of service, last updated 5 August 2026, state that no company name, registered address or place of incorporation appears on the site and that governing law is 'not chosen anywhere' (15 October 2026)gandr.ai ↗
Is Gandr legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Domain registered
- Feb 4, 2026Registry record (RDAP), registrar Porkbun LLC, looked up Sep 11, 2026
- TLS certificate
- Issued by Google Trust Services, valid until Nov 28, 2026Seen in our own connection to the site on Oct 15, 2026
- Last checked
- Oct 15, 2026The site answered
- Operated by
- GandrOur research, card verified Oct 15, 2026
Pros & cons of Gandr
Gandr's stream-based pricing is the notable design choice here: instead of metering every character or minute, it charges for reserved concurrent capacity, which favors applications with heavy, unpredictable usage on a fixed number of simultaneous calls, like a phone-based voice agent. That same model works against low-volume or bursty use cases, where a per-character competitor could end up cheaper.
Pros
- Voice cloning needs only about ten seconds of reference audio with no separate training step
- 146ms first-audio-byte latency is fast enough for live conversational use
- Stream-based pricing gives predictable cost regardless of how heavily a stream is used
- Direct integration with common voice-agent frameworks such as LiveKit, Pipecat, and Vapi
Cons
- Stream-based billing can cost more than usage-based competitors for light or unpredictable workloads
- It is API-first infrastructure rather than a consumer app, so it requires development work to use
- The published terms of service still contain unfilled placeholders: no legal entity, address or governing law is named
Best for
- Voice-agent builders needing predictable per-stream cost regardless of usage volume
- Applications requiring low end-to-end audio latency
- Teams already using LiveKit, Pipecat, or Vapi for voice infrastructure
Not for
- Low-volume or occasional text-to-speech use, where per-character pricing from a competitor may cost less
- Teams wanting a no-code consumer app rather than an API
Watch out: The terms page (dated 5 August 2026) literally says no company name, registered address or place of incorporation appears on the site and that governing law has not been chosen; a compliance review for a HIPAA or GDPR workload will stall on that until Gandr names a counterparty.
Alternatives to Gandr
Gandr head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.
Also used for
AI tools by profession
Text-to-Speech is one of the jobs on these pages: