workflow

workflow is the orchestration layer for your natural language-driven tasks. It allows you to build complex workflows, which can be triggered by a variety of events. For instance, you can use workflow to grab data from a 3rd party API, build an NLP in refinery and gates, and then use the results in any further step of the pipeline.

Structure of this documentation

The documentation is structured by features of workflow. If a feature is accessible in some extent via API, there is a small API badge next to it. In this case, you can find the API specs at the bottom of the page.

Featured use cases

Combined with refinery and gates, workflow is quite powerful for a set of featured use cases. To get a better feeling, we added some sample use cases in here:

• Building training data for NLP
• Automating emails
• Webscraping
• NLP ETL

Features of workflow

workflow comes with the following features.

Drag-and-drop-editor

Simple drag-and-drop interface to build your workflows, connected to catalogue of nodes with either no-code or programmable interfaces. This is as intuitive as it gets.

Completing the stack

workflow stands on the shoulders of refinery and gates. This means that workflow is capable of handling the most complex NLP applications (and of course also the simpler ones), while still being easy to use.

Integrations

workflow offers native integrations to e.g. Google workspace applications, Slack or other collaboration channels, and further offers an API and Webhook links. We are continuously adding new integrations.

Data collection

Data is being stored in stores, such that you have direct access to your data source integrations. For instance, you can just export all emails from your inbox into refinery via the GMail integration, and then start building your NLP automations.

Realtime and batch

workflow is designed to work out of the box for the use cases you want to implement, whether it is realtime or batch processing. You can run operational tasks in realtime, and batch processing can be used for e.g. data analysis. Alternatively, you can synchronize refinery in a batch-job, such that your training data is always up-to-date. You will quickly realize: This is ETL for NLP - and it is designed to fit your ideas.