This is how I do the same thing using Markdown and wiki-links for lists and tables…
You and a few others have earlier asked about how I use Obsidian, Foam, and Markdown as my research tools, and this is a classic example. In Markdown, I can use a Python script, an LLM with an MCP, or simply internal Regex search-and-replace tools in VS Code or Obsidian to handle the entire process semantically.
When lists or tables contain hundreds or thousands of objects, manually creating links within a Gramps note becomes challenging. This is one of the main reasons I started using Markdown: to be able to create links in source documents that semi-automatically link to existing object notes within my Markdown folder structure (also known as a “Vault” in Obsidian), and to be able to view these connections visually. In the visual network graph in Obsidian or Foam, you can apply colors, sizes, and other meta-attributes to nodes and edges using graph filters. For example, all existing objects can be colored green, while non-existent objects (unresolved links) appear gray or red, and “orphan” notes get their own distinct color. This makes it incredibly easy to spot if you have actually discovered something new, uncovered hidden relationships between objects, or found a new valuable source.
Just as you link to Gramps Objects (like people), I link to objects in the form of Markdown notes—one main note for each “object”. These notes can easily be added to Gramps as media files, including both the object notes and the source notes.
The example you posted is also a perfect use case for how an LLM or VLM can be extremely useful in Gramps: transcribing a document into a Gramps note, searching the database via MCP, suggesting potential objects as targets for internal linking, and formatting the note strictly according to a specific template.
Note: This information is shared for context and workflow illustration only. It is not intended to pivot this specific thread into a discussion about LLMs or external markdown tools.
Author’s Note: This response was translated from my original Norwegian thoughts, structured, and edited for English flow and technical clarity by Google AI (Gemini 3.5), operating strictly under my direction and inputs.