Context sources in Maxim allow you to expose your RAG pipeline via a simple endpoint irrespective of the complex steps within it. This context source can then be linked as a variable in your prompt or endpoints and selected for evaluation.

​
Bring Your RAG via API Endpoint

Connect your existing RAG system to Maxim by exposing it via an API endpoint.
1

Create a new context source

You can create a new context source by clicking on the left navigation and then the plus icon.
2

Name your context source

Provide a name for the context source.
3

Configure API endpoint

Enter the API endpoint. When we pass a query to this endpoint, it should return the retrieved context for that query as would be the real world scenario.
4

Set up request configuration

Add necessary headers and parameters to your request and save the context source
5

Implement in prompts and endpoints

You can now use this in prompts and endpoints.

​
Ingest Files as Context Source

Maxim allows you to upload various document formats and automatically creates embeddings for use as context in your applications.
1

Navigate to Context Sources

Navigate to the Context Sources section in your workspace and click on the Plus icon.
2

Select Files source type

Select “Files” as your context source type.
3

Upload your documents

Drag and drop your files or click “Browse files” and select files from your system. Supported formats:
  • PDF documents (.pdf)
  • Word documents (.docx)
  • CSV files (.csv)
  • Text files (.txt)
  • Markdown files (.md)
4

Wait for processing

Wait for the files to upload and for embeddings to be generated. The status indicator will show progress.
5

Use your context source

Once processing is complete, your files are ready to be used as context. You can now reference this context source in agent simulation.

​
Evaluate Your Context

Context sources in Maxim allow you to expose your RAG pipeline via a simple endpoint irrespective of the complex steps within it. This context source can then be linked as a variable in your prompt or workflow and selected for evaluation here.