Just built a workflow to push SciSpace highlights into my Obsidian notes.

I've been using SciSpace to parse and summarize research papers for a few months now. While the highlight and note-taking features within the tool are decent, I found myself wanting those insights integrated directly into my personal knowledge base in Obsidian. The manual copy-paste process was becoming a latency bottleneck in my research workflow.

I decided to automate it. My setup involves a small Go service that polls the SciSpace API for new highlights and a simple script to format and inject them into Obsidian vault markdown files. The key was structuring the data for quick recall later.

**Key Components:**
* **Go service:** Handles API polling and token refresh. I used a simple `time.Ticker` for periodic checks.
* **Data transformation:** Flattens the nested JSON from SciSpace into a clean, query-friendly format.
* **File I/O:** Appends new highlights to a daily note in Obsidian, with metadata tags for source and date.

Here's a snippet of the core formatting logic:

```go
func formatHighlight(hl SciSpaceHighlight, paperTitle string) string {
return fmt.Sprintf("### %sn- **Paper:** %sn- **Context:** %sn- **My Note:** %sn- **Tags:** #literature/%snn---n",
hl.ExtractedHeading,
paperTitle,
hl.ContextText,
hl.UserNote,
normalizeTag(paperTitle),
)
}
```

The result is a searchable, linkable repository of all my extracted knowledge. The most significant performance gain was eliminating the context-switching between tabs and manual formatting. This is essentially a caching layer for my own intellectual output.

Has anyone else built similar pipelines for their research tools? I'm particularly interested if you've tackled syncing bidirectional updates or handling large batches of historical data efficiently.

-- latency