If you use talkkonnect, please let us know your use case by sending us pictures, and please also STAR the talkkonnect/talkkonnect repo on github.com!

• What is talKKonnect?
• Installation / Getting Started
• Functionality and Configurability
• Using the Terminal CLI
• Voice Targets (Whisper and Shout)
• Multicast to IP Speakers and SIP Phones
• HTTP API Specification / Building Your Own UI
• Configuring and Running talKKonnect
• Extra Multimedia Features (IP-Speaker)
• Software Configurable Features
• Additional Optional Hardware and Precautions
• Common Information for all the Pre-Made Images For Various Hardware Configurations
• Why Was talKKonnect created?
• Questions & Contributing
• License

talKKonnect is an open-source customizable, headless, self-contained Mumble Push to Talk ( PTT) client.

It was designed for Linux single-board computers (SBCs) such as the Raspberry Pi and Orange Pi. It works equally well on any reasonably modern Linux distribution.

talKKonnect offers a flexible form factor with an LCD/OLED display, channel and volume control, making it ideal for group communication scenarios. Common use cases include amateur radio enthusiasts, ad-hoc group communications, and replacing expensive commercial intercom systems. Read more about talKKonnect as an appliance and some potential use-cases.

Diving right in? talKKonnect works on a variety of devices and form factors, and even provides pre-built images for use on common Raspberry Pi or similar single board computer (SBC) architectures.

• General Installation instructions
• Raspberry Pi / Other Pre-made Image Instructions
• Configuration and Running

Because talKKonnect was originally created as software to power hardware-based IP communication devices, it has extensive optional functionality and configurability for several common hardware devices. See a full overview of talKKonnect configuration and support details

When talKKonnect runs in a terminal it keeps a command prompt pinned to the bottom row while the log scrolls above it. Everything the device can do is reachable from there: single keys for the common actions, typed commands with Tab completion for the rest, and the same prompt remotely over SSH. Read the terminal CLI reference

A voice target sends your transmitted audio to specific users or a specific channel instead of the channel you are joined to — whispering and shouting in Mumble terms. It is how one device pages a single zone rather than everybody, and the vt command sets targets up without hand-editing the XML. Read about voice targets and the vt command

talKKonnect can re-transmit the audio it receives from Mumble to a multicast group as RTP, so hardware IP PA speakers (CyberData, Algo, Barix, Advanced Network Devices) and SIP desk phones such as Yealink can hear a Mumble channel without being Mumble clients. This turns a talKKonnect node into a gateway between a Mumble channel and an existing paging system. Read about multicast output

talKKonnect exposes two HTTP interfaces: a JSON telemetry endpoint (/uistatus) that publishes the complete live state of the client, and a control API (/?command=...) that executes remote control commands. Together they let you build any user interface you like, in any language, with no changes to talKKonnect itself.

Read the full HTTP API specification — it documents every endpoint, field, command and status code, and includes build guides for the two interfaces people most often want:

• A framebuffer screen — a full-screen status display drawn straight to a Linux framebuffer, HDMI monitor or SPI TFT panel on the talkkonnect device itself, with far more room than a 4x20 character LCD. See the guide
• A web UI — a browser dashboard showing live state and sending control commands. See the guide

Both are small, well-specified projects that an AI programming agent (such as Claude Code) can write almost end to end. The specification includes ready-to-use starting prompts, guidance on iterating with an agent, and a review checklist covering the mistakes that actually show up in generated code. See the guide

If you build a UI this way, please share it with us — we would like to feature community-built interfaces.

This project was created by Suvir Kumar as a fork of talkiepi by Daniel Chote which was, in turn, a fork of barnard a text based mumble client. talKKonnect was developed using golang and based on gumble library by Tim Cooper. Read the full history and background of talKKonnect

See a video explanation of the history and reasons for creating talkkonnect

We invite interested individuals to provide feedback and improvements to the project. You can help with creating youtube videos, documentation, programming, testing and/or feedback of your user experience.

To speak to us, connect with a standard mumble client (Android/iPhone/Windows/Linux), or with talkkonnect itself, to our community server at mumble.talkkonnect.com port 64738. Use any unique username; no password is required. We are standing by usually on the HAM-CB channel.

Currently we do not have a WIKI, so send feedback to suvir@talkkonnect.com or open an Issue on GitHub.

Please visit our blog for updates on the project, our github for the latest source code, and our facebook page for future updates and information.

Thank you all for your kind feedback sent along with some pictures and use cases for talkkonnect.

talKKonnect is open source and available under the Mozilla Public License 2.0.

suvir@talkkonnect.com Updated 29/07/2026. talkkonnect version 4.18.01 is the latest release as of this writing.