In the context of HoneyHive, a project refers to a collection of prompts, models, datasets, evaluation runs and metrics that work together to accomplish a specific task for which you’re using an LLM. For example, a SQL code generator will be considered a seperate Project vs an email generator within the same application.

It is best to organize each project within your workspace as a separate task for which you’re using large language models. This allows you to standardize data structures and schema for each individual project but allows the flexibility to store multiple prompt variants within each Project space.

​
Create a new project

  1. Log in to HoneyHive via our authentication client.
  2. Create a new project by clicking on the Add Project button.
  3. Give your project a name (and optionally a type and description as well) and click Create Project.

projects

In instances where you’re chaining multiple LLM & tool calls together, we recommend instrumenting each component of a chain inside the same project within HoneyHive.
Tracing LangChain PipelinesOverview