End user documentation

See how we use weak supervision as a method to cleanse labels.

See how versioning in cognition works.

See how to upload data into cognition.

Change the attribute visibility in the settings page.

See how uploading pdfs while chatting can be used.

Put your users into teams to manage access for projects.

See which roles are supported in cognition

This is the knowledge base a GenAI assistant can access.

Use the record IDE to quickly analyze a single record structure.

See how to create projects in refinery and upload data

Use refinery to identify outliers in your dataset.

refinery can be used to label with multiple users and see the inter-annotator-agreement

You can use refinery to analyze and monitor your data.

See how you can manually label data in refinery

See how you can manage roles in refinery

See how to use macros (repeatable action sequences)

See how you can create labeling tasks to label your data

See how you can use heuristics in refinery to automate data labeling

See how you can evaluate heuristics in refinery.

Environment variables are required to connect to e.g. cloud LLMs.

Add embeddings directly from OpenAI, Azure, Hugging Face, Cohere or other embeddings.

On Cognition, there is a data processing pipeline known as ETL (Extract, Transform, Load). This pipeline handles the processing of data in the form of PDF files.

Data management is a key feature of refinery. See how we enable it here.

See how you can export the data in your refinery project

Download and analyse conversations from a GenAI agent.

See how you can set relevant data settings and labeling tasks for your project.

See how you can add comments to work better with colleagues or to set reminders for your own process.

Use the pre-built chat UI to get your system up and running as fast as possible.

The refinery platform includes a bricks integration feature designed to simplify the process of identifying and importing available brick modules for specific use cases. Users are prompted by the integrator to either assign values to user-defined variables or modify the default variable values as necessary. Once these adjustments are made, the code is prepared for execution within the refinery environment.

See how your GenAI agent performs and how you can improve it.

See how to configure the action steps an GenAI agent can do.

See where to get started in cognition

Refinery provides a feature where users can enhance their existing data by computing new attributes. This is achieved by writing Python code that manipulates the current data records, potentially utilizing external APIs, to produce and return new values for these added attributes.