We’ve just released full multi-language AI agent extension support for the TEN Framework, an open-source realtime multimodal AI agent framework designed for cloud-edge compute fusion.
This update includes SDKs and testing frameworks for writing extensions in:
* C++
* Go
* Python
* Node.js (TypeScript/JavaScript)
We also published a comprehensive guide that walks through the entire lifecycle of extension development:
The guide covers a full self-contained development loop:
* Create: Scaffold a new TEN extension in your preferred language
* Install dependencies: Language-specific setup for build/test/runtime environments
* Develop: Write business logic using a stable interface definition
* Test: Use our unified extension testing framework to validate behavior
* Debug: Inspect runtime execution and cross-language interactions
Whether you're building a new extension or customizing existing ones, this update is designed to make the process faster, easier, and more reliable.
Extensions can run in-process with the AI agent or be distributed across cloud and edge environments - all under the same multimodal runtime. TEN is designed to allow realtime coordination across perception, dialogue, planning, and actuation, using the developer’s language of choice.
We’d love feedback from the HN community, especially around multi-language architecture, testing flows, and edge deployment strategies. Contributions and issues welcome!
This update includes SDKs and testing frameworks for writing extensions in:
* C++
* Go
* Python
* Node.js (TypeScript/JavaScript)
We also published a comprehensive guide that walks through the entire lifecycle of extension development:
https://theten.ai/docs/ten_framework/development/how_to_deve...
The guide covers a full self-contained development loop:
* Create: Scaffold a new TEN extension in your preferred language
* Install dependencies: Language-specific setup for build/test/runtime environments
* Develop: Write business logic using a stable interface definition
* Test: Use our unified extension testing framework to validate behavior
* Debug: Inspect runtime execution and cross-language interactions
Whether you're building a new extension or customizing existing ones, this update is designed to make the process faster, easier, and more reliable.
Extensions can run in-process with the AI agent or be distributed across cloud and edge environments - all under the same multimodal runtime. TEN is designed to allow realtime coordination across perception, dialogue, planning, and actuation, using the developer’s language of choice.
We’d love feedback from the HN community, especially around multi-language architecture, testing flows, and edge deployment strategies. Contributions and issues welcome!
GitHub: https://github.com/TEN-framework/ten-framework
Docs: https://theten.ai/docs/