Promote Sign in Sign up

Text to speech 2026 Rank: #391

Gandr ✓

Unlimited text-to-speech for AI voice agents.

Source rating 4.8 (5) 634 views 19 saves 8 comments Free Trial
API TAAFT Code Text To Speech
More categories (4)
Voice Cloning AI Voice Agents Voice Agents Text to speech
Visit website Sign in to save

Trial; then from $10.00/mo

Privacy Terms

Verdict

A quick read before you dig into the full page.

Best if
Unlimited text-to-speech for AI voice agents
Skip if
Limited language support (23)
Pick this when
Cheapest TTS in the market

Considering ElevenLabs instead? Compare side by side

Overview

Gandr is a text to speech and voice cloning API for voice agents. Audio is streamed back in chunks as it is generated rather than delivered after the whole clip finishes rendering, so a phone agent can begin playback while the rest is still being produced.

It reads the parts that usually break. Order numbers, phone numbers and addresses are read the way a person reads them. On the standard word error rate test it scored 1.982 percent against a 2.171 percent human reference passed through the same scorer.

One cloned voice speaks 23 languages, including languages the reference recording never contained. Cloning takes a ten second reference clip, there is no per voice fee, and the clone is cached for reuse.

Streaming runs over WebSocket for agents, with HTTP for narration and dubbing. It suits audiobooks, dubbing, video, e-learning, games, podcasts and support lines.

No user data enters any training set. Every synthesized clip carries an inaudible watermark, so generated audio stays identifiable downstream.

Pricing is published in full. A free key covers 50,000 tokens. Paid plans start at $10 a month, and a flat unlimited unmetered stream is $150 a month for teams that would rather stop counting.
  • Unlimited Unmetered Stream Plans
  • Websocket Streaming For Agents
  • Instant Voice Cloning From A Reference Clip
  • 23 Languages
  • Inaudible Watermark On Every Render
  • Rest + Sse + Websocket Apis
  • Per-customer Pronunciation Dictionaries
  • Mcp Server And Python/js/pipecat Sdks
  • No Customer Data In Any Training Set
Gandr website

Access & connections

Where you can use Gandr, and what it plugs into.

Platforms

Web

Integrations

MCP

API

Public API available

Browse other tools with an API

Capabilities

What this tool accepts, produces, and is strongest at.

Input

  • API
  • Text
  • Audio

Output

  • Audio
  • API

Highlights

  • API
  • MCP
  • Agents
  • Use code TAAFT
  • 10% OFF

Supported features

  • API
  • MCP
  • Agents
  • Speech API

Pros & cons

Pros

  • Cheapest TTS in the market
  • Voice Cloning and Voice changer
  • SSE Support
  • Unlimited text-to-speech service
  • Supports 23 languages
  • Functional in multiple applications
  • WebSocket for agent integration
  • HTTP for narration and dubbing
Show 11 more
  • Rapid voice-cloning feature
  • Creation of clone from ten seconds reference
  • No per-voice fee
  • Cached voice clones for reuse
  • Inaudible audio watermarking
  • Guaranteed user privacy protection
  • Transparent and predictable pricing
  • Concurrent streams based pricing
  • Operational redundancy for reliability
  • Supports audiobooks to games applications
  • Diverse user need fulfillment

Cons

  • 10 seconds minimum for cloning
  • Inaudible watermarking mandatory
  • Pricing based on concurrency
  • Limited language support (23)
  • Redundancy might cause complexity
  • No personalized training sets

Compare alternatives

Similar tools in the catalog — with a clear reason to open each one.

Why look Typically cheaper

Full comparison

No pricing

Full comparison

Why look Better for writing

Full comparison

Why look Fully free option

Full comparison

Why look Typically cheaper

Full comparison

from $19.99/mo

Full comparison

Releases 1

Initial release August 10, 2026 Current
Initial release of Gandr.

Imported reviews 4

5★
4
4★
1
3★
0
2★
0
1★
0
Joseph Kasieen Aug 10, 2026 v@ initial release

I think it's great. I had a major problem with scaling due to TTS, so this is helping erase that. If I do encounter a problem, I will let you guys know.

Hameed A Aug 10, 2026 v@ initial release

I finally sleep easy knowing Gandr’s flat-rate pricing won't bankrupt us if we go viral, and the ultra-low latency makes our voice agent sound completely human.

Johan Aug 10, 2026 v@ initial release

With phone agents relying on clear speech, how much control do teams have over pronunciation for things like product names, addresses and industry-specific terms?

Mattias Aug 10, 2026 v@ initial release

Gandr is a solid option for voice agents, especially with WebSocket streaming and the quick voice cloning setup. I also like the focus on pronunciation for things like phone numbers, dates and addresses, since those details often sound awkward with TTS. The 23-language support is another nice addition.

FAQ 12

What is Gandr?
Gandr is a platform specializing in text-to-speech and voice-cloning that provides unlimited service for voice agents. It offers support for 23 languages and is designed to serve a variety of applications such as audiobooks, dubbing, videos, e-learning, games, podcasts, and voice agents.
What is the purpose of Gandr's voice-cloning feature?
The purpose of Gandr's voice-cloning feature is to create a functional clone of a voice from a ten seconds reference clip. This feature is particularly useful in applications such as AI voice agents, narration, dubbing, and other derived uses. Once created, the voice clone is cached for future use, thus saving time and reducing the need for repetitive cloning of the same voice.
How does Gandr's per-stream pricing work?
Gandr's per-stream pricing model is based on concurrent streams. This means that the cost is determined by the number of streams running at the same time, providing unlimited service within those streams. This makes Gandr's pricing model simple and predictable, with costs remaining consistent regardless of the volume of usage within each stream.
How can I integrate Gandr's text-to-speech functionality into my current set-up?
To integrate Gandr's text-to-speech functionality into your current setup, you can utilize WebSocket for agents and HTTP for narration and dubbing. This allows for a smooth integration process and compatibility with various existing setups.
What are the 23 languages supported by Gandr?
Gandr supports 23 languages which include English, Spanish, French, German, Portuguese, Arabic, and Chinese among others. Specific languages are not explicitly listed on their website.
What applications does Gandr cater to?
Gandr caters to a wide array of applications such as the creation of audiobooks, dubbing for movies or animations, video narrations, e-learning modules, gaming dialogues, podcasts, as well as providing voice for AI voice agents in various languages.
How does Gandr ensure user data privacy?
Gandr ensures user data privacy by employing stringent measures to prevent user data from entering any training set. This secures the privacy of user information and their identities, since any data sent over Gandr's platform is not used for machine learning or similar procedures.
Can Gandr clone any voice?
Gandr can clone any voice using a ten-second reference clip. The rapid voice-cloning feature of Gandr allows the creation of a working clone from just a short reference clip, without any additional per-voice fee.
What is the meaning of a 'concurrent stream' in Gandr's pricing model?
A 'concurrent stream' in Gandr's pricing model refers to the number of text-to-speech or voice-cloning streams running at the same time under the service. Regardless of usage volume within each stream, the cost is based solely on the number of concurrent streams.
What does the inaudible watermark in Gandr's audio clips do?
The inaudible watermark in Gandr's audio clips serves as a unique digital signature, verifying the origin of the synthesized audio clips and ensuring integrity and authentication of the output from Gandr's platform.
Why does Gandr use WebSocket for agents?
Gandr uses WebSocket for agents to facilitate real-time communication between the server and the client. WebSocket provides faster, bi-directional communication, making it suitable for highly interactive applications such as voice agents.
How quickly can Gandr clone a voice?
Gandr can clone a voice quickly, creating a functional clone from just a ten seconds reference clip. The exact timeframe is not specified on their website.
Gandr Trial; then from $10.00/mo
Visit website