TCP/IP
TCP/IP protocol
Connect your application directly to the speechServer engine over a TCP/IP communication protocol.
ReadSpeaker speechServer
On-premise text-to-speech server for your own network
Run ReadSpeaker's text-to-speech engine on your own Windows Server or Linux machines. Your applications request speech over TCP/IP, the command-line tool, the REST API, or the C, Java, and .NET SDKs.
- 150+ AI voices
- 50+ languages
- Self-hosted on Windows Server or Linux
INSIDE speechServer
A text-to-speech server you host yourself
ReadSpeaker speechServer is a client-server, network-based text-to-speech server. You install it on your own infrastructure, and your applications request speech synthesis over the network. It suits teams that need text-to-speech inside their own environment, from virtual assistants and notification platforms to educational applications and other server-based systems that generate audio on demand.
HOW IT CONNECTS
Four ways to integrate your application
Your application talks to speechServer the way that fits your stack. Connect over a direct network protocol, the command-line tool, a web API, or a client SDK in your own language.
CLI
Command-line tool
Drive speech synthesis from scripts and server processes with the included command-line tool.
REST
REST API
Call speechServer over a REST API when it runs behind a CGI-enabled web server such as Apache.
SDK
C, Java, and .NET SDKs
Integrate with client-side SDKs for C, Java, and .NET that handle communication between your application and speechServer.
# direct socket connection
connect speechserver:5555
synth "Welcome"speechserver --voice Emma \
--text "Hello" --out out.wavPOST /synthesize
{ "voice": "Emma", "text": "Hi" }// .NET client SDK
var client = new SpeechServer();
client.Synthesize("Welcome");Code samples are illustrative. See the user documentation for the exact API syntax.
RUN IT YOUR WAY
Runs on Windows Server and Linux
Install speechServer on the platform your operations already run.
Windows Server
Install speechServer on your Windows Server machines, alongside the services your operations already run.
Linux
Install speechServer on the Linux distributions your infrastructure is built on, from bare metal to virtual machines.
AUDIO OUTPUT
Stream the audio format your system needs
speechServer streams audio back to your application in the format your pipeline expects, from telephony-grade encodings to standard media formats.
PCM
PCM Wave
A-law
μ-law
ADPCM
mp3
OGG
VOICES
150+ AI voices in 50+ languages
speechServer uses ReadSpeaker's catalog of more than 150 AI voices across over 50 languages, with new voices and languages added regularly. Each licensed language includes a user dictionary so you can fine-tune pronunciation for domain-specific terms.
Global coverage
150+ AI voices in 50+ languages
Available languages
- FrançaisMarie · fr-FR
- English (UK)Alice · en-GB
- English (US)Kayla · en-US
- DeutschEmil · de-DE
- EspañolIsabel · es-ES
- NederlandsSara · nl-NL
- SvenskaMaja · sv-SE
- ItalianoGaia · it-IT
- Português (BR)Julia · pt-BR
- PolskiAneta · pl-PL
- العربيةAmir · ar-XA
- 日本語Akira · ja-JP
- 中文Lin · zh-CN
- 한국어Jihun · ko-KR
- DanskMette · da-DK
- SuomiElina · fi-FI
- NorskErik · nb-NO
FINE CONTROL
Control how every word is spoken
Shape the delivery of your text down to the word, with the controls your content already carries.
SSML voice and language switch
Switch to another voice, or another language, mid-text using SSML instructions in your input. Adjust speaking rate, pitch, and volume as needed.
User dictionary and IPA
Define custom pronunciations for words and patterns with per-language user dictionaries and IPA transcription input, so domain terms read correctly.
Audio clip insertion
Insert references to audio clips in your input text. speechServer places the referenced audio at the right position in the streamed output.
Mark information
Receive mark information alongside the audio for event triggers at text positions, device and interface synchronization, and text highlighting aligned with playback.
DEPLOYMENT
Your text-to-speech stays in your network
speechServer runs entirely on your own servers, so text and audio stay inside your network.
Installed, licensed and operated by you
speechServer runs entirely on your own servers, so text and audio stay inside your network. It installs with the speechServer installer, and any licensed ReadSpeaker voice can be added. Use is governed by a license file that sets your voices, concurrent ports, synthesis speed, number of servers, and term.
- Installation and implementation support is included from the ReadSpeaker Support Team.
speechServer vs OPEN-SOURCE
A supported alternative to do-it-yourself text-to-speech
Open-source projects can run text-to-speech locally, but you maintain the engine, the voices, and the integration yourself. speechServer gives you a commercial, supported server with a maintained voice catalog.
| ReadSpeaker speechServer | Open-source do-it-yourself | |
|---|---|---|
| Type | Commercial product, supported | Free open-source, community-maintained |
| Voice catalog | 150+ AI voices, 50+ languages | Bring-your-own, varies |
| Integration | TCP/IP, command line, REST, C/Java/.NET SDKs | Do-it-yourself wiring |
| Support | ReadSpeaker Support Team, installation and implementation | Community |
| Licensing | License file with defined terms | Project license, no SLA |
| Voice updates | Added regularly | Do-it-yourself |
FAQ
Questions about speechServer
ReadSpeaker speechServer is an on-premise, network-based text-to-speech server. You host it on your own Windows Server or Linux machines, and your applications request speech synthesis over the network. It provides the ReadSpeaker text-to-speech engine, a server API, and tools as a complete server-side runtime.
speechServer is self-hosted, not cloud-based: it runs on your own Windows Server or Linux machines, and your text and the audio it generates stay inside your own network rather than passing through a third-party service. If you prefer a managed cloud API instead, ReadSpeaker speechCloud API is the hosted option.
Client applications connect to speechServer over a TCP/IP communication protocol, a command-line tool, or a REST API when it runs behind a CGI-enabled web server such as Apache. Client-side SDKs for C, Java, and .NET are also provided.
speechServer provides client-side SDKs for C, Java, and .NET that handle the communication between your application and the server, so you integrate speech synthesis in your own code rather than wiring the protocol by hand. Supported development languages are C and C++, Java, and C# / .NET on Windows.
speechServer runs on Windows Server and on Linux. Ask our team for the current list of supported and tested versions for your environment.
speechServer outputs 16-bit linear PCM, PCM Wave variants, 8-bit A-law and μ-law PCM, 4-bit Dialogic ADPCM, mp3 through the LAME package, and OGG. This range covers both telephony-grade and standard media playback.
speechServer supports ReadSpeaker's catalog of more than 150 AI voices across over 50 languages. New voices and languages are added regularly. Each licensed language includes a user dictionary and IPA support for custom pronunciations.
speechServer is a commercial, supported product with a maintained catalog of 150+ AI voices, four integration paths, and installation support from the ReadSpeaker Support Team. Open-source projects can run locally, but you maintain the engine, voices, and integration yourself, without an SLA.
speechServer integrates directly through TCP/IP, command line, REST, or C, Java, and .NET SDKs, for general server-based applications. ReadSpeaker speechServer MRCP uses the MRCP protocol for standards-based telephony and IVR platforms.
speechServer use is governed by a license file. It sets the licensed voices, the number of concurrent text-to-speech ports, the permitted synthesis speed, the number of servers covered, and the license term. Concurrency is limited only by your licensed ports and available system resources.