NeuDecide

Use it to turn spoken commands into structured calls to tools, such as controlling a robot vacuum.

Availability

Hugging Face repository; install with pip install neudecide; access to model files requires accepting conditions

Pricing model

Not established

Deployment

Runs locally via ONNX Runtime; no GPU or internet connection needed after download

Privacy

Can run locally; no internet connection required after download.

Licence

Apache 2.0

• NeuDecide converts speech directly into function calls without an intermediate transcript. ^[1]

  NeuDecide converts speech directly into function calls.

• Tool definitions are passed as JSON schemas at inference time. ^[1]

  Tools are passed in as JSON schemas at inference time, so you change what the model can do by changing the JSON, not by retraining.

• The model runs on a single thread via ONNX Runtime and needs no GPU or internet connection once downloaded. ^[1]

  It's 43 MB, can run on a single thread via ONNX Runtime, and needs no GPU or internet connection once downloaded.

• The package can be installed with pip install neudecide and requires Python 3.10–3.13. ^[1]
• Access to the Hugging Face model files requires agreeing to share account information and accepting repository conditions. ^[1]

  You need to agree to share your Hugging Face information to access this model

• The model supports English only, accepts audio up to 30 seconds, and is recommended for use with at most 10 tools. ^[1]

• Turn a prompt into an editable slide, poster, or dashboard Weekly brief 2026-W41
• Extract document layout and chart data, not just text Daily brief 2026-10-08