Upload your data manually (CSV, text), via our API or command-line interface. HumanFirst also integrates directly with various unlabeled and labeled data sources.

HumanFirst preserves the context of utterances, and links back every sentence to its conversation.

The data pipeline jobs and artifacts are exposed via our APIs and CLI - you can upload, manage, export, and launch training or evaluation reports with a single command (or line of code).

Your data is automatically projected to a latent space, which powers the similarity-search and other real-time features / APIs. HumanFirst indexes every unlabeled utterance against its trained NLU model.

HumanFirst is made for scale and can process and index millions of data points making them available for real-time query within the platform.

Plug in a 3rd-party NLU provider (we currently support Rasa, and will be adding DialogFlow and Watson soon) or your own custom NLU.

HumanFirst preserves the context of utterances, and links back every sentence to its conversation.

The data pipeline jobs and artifacts are exposed via our APIs and CLI - you can upload, manage, export, and launch training or evaluation reports with a single command (or line of code).

Your data is automatically projected to a latent space, which powers the similarity-search and other real-time features / APIs. HumanFirst indexes every unlabeled utterance against its trained NLU model.

HumanFirst is made for scale and can process and index millions of data points making them available for real-time query within the platform.

Plug in a 3rd-party NLU provider (we currently support Rasa, and will be adding DialogFlow and Watson soon) or your own custom NLU.

HumanFirst preserves the context of utterances, and links back every sentence to its conversation.

The data pipeline jobs and artifacts are exposed via our APIs and CLI - you can upload, manage, export, and launch training or evaluation reports with a single command (or line of code).

Your data is automatically projected to a latent space, which powers the similarity-search and other real-time features / APIs. HumanFirst indexes every unlabeled utterance against its trained NLU model.

HumanFirst is made for scale and can process and index millions of data points making them available for real-time query within the platform.

Plug in a 3rd-party NLU provider (we currently support Rasa, and will be adding DialogFlow and Watson soon) or your own custom NLU.

HumanFirst preserves the context of utterances, and links back every sentence to its conversation.

The data pipeline jobs and artifacts are exposed via our APIs and CLI - you can upload, manage, export, and launch training or evaluation reports with a single command (or line of code).

Your data is automatically projected to a latent space, which powers the similarity-search and other real-time features / APIs. HumanFirst indexes every unlabeled utterance against its trained NLU model.

HumanFirst is made for scale and can process and index millions of data points making them available for real-time query within the platform.

Plug in a 3rd-party NLU provider (we currently support Rasa, and will be adding DialogFlow and Watson soon) or your own custom NLU.