Uploaded November 2026 | Updated November 2026, 4 days ago
Synaptics demonstrates its SL2619 Coral Board, a development platform for running AI models locally at the edge. The demo powers an offline language model with a hand crank, then uses speech recognition and text-to-speech for a spoken interaction. synaptics.com
The system combines a Liquid AI language model with speech components identified in the interview as “Bipro” and “Munchine.” The demo shows a voice interface operating without a battery, while the exact power figures and component names are not fully clear in the recording.
The board is described as co-developed with Google Research and Green, and available to order. The interview identifies an NPU and a microSD connector; it also describes a path to cloud connectivity when needed, while keeping as much processing as possible on-device.
Synaptics says development resources are available through its developer site and GitHub. The open-source materials described include the SDK, AI toolchain, model-compilation tools, and NPU runtime components, with support for developers adapting or training models for the platform.
The demo illustrates a compact edge-AI workflow spanning local inference, speech input and output, model deployment, and optional network access. The interview was filmed at electronica 2026.
Synaptics demonstrates its SL2619 Coral Board, a development platform for running AI models locally at the edge. The demo powers an offline language model with a hand crank, then uses speech recognition and text-to-speech for a spoken interaction. synaptics.com
The system combines a Liquid AI language model with speech components identified in the interview as “Bipro” and “Munchine.” The demo shows a voice interface operating without a battery, while the exact power figures and component names are not fully clear in the recording.
The board is described as co-developed with Google Research and Green, and available to order. The interview identifies an NPU and a microSD connector; it also describes a path to cloud connectivity when needed, while keeping as much processing as possible on-device.
Synaptics says development resources are available through its developer site and GitHub. The open-source materials described include the SDK, AI toolchain, model-compilation tools, and NPU runtime components, with support for developers adapting or training models for the platform.
The demo illustrates a compact edge-AI workflow spanning local inference, speech input and output, model deployment, and optional network access. The interview was filmed at electronica 2026.










