Skip to main content
rogier
Enthusiast
July 26, 2026
Solved

Extracting new message content from the API (existing solutions?)

  • July 26, 2026
  • 1 reply
  • 52 views

Setup: I sync messages from the Core API and process them downstream. For that I need just the new content of each message, not the quoted thread history, signatures or disclaimers.

What the API gives me:
- `body` - full HTML, including everything
- `text` - plaintext rendering, also including everything
- `blurb` - quote-free, but truncated, so usable as a signal at best

So I'm parsing the HTML myself, and running into the usual wall: no standard for marking quoted content, every client does it differently, unclosed tags, tables used for layout. Forwards are the worst case, some clients wrap the entire forwarded message in a single blockquote, so naive stripping leaves an empty body.

Front clearly handles this internally (collapsed quotes in the UI, the `front-quoted` class, rules that skip quoted text when matching conditions).

1. Is that exposed anywhere I've overlooked, an endpoint, field, parameter, or a paid add-on? Happy to pay for it.
2. If not, what are you all using? Interested in what holds up on real volume, and whether anyone has solved this for non-English mail (Dutch, in my case).

Any pointers appreciated.

Best answer by raymond_ching

Hi Rogier,
With regards to parsing out quoted content or anything similar, we’ve built a JS version of the Talon quote-detection framework that you can see here: https://github.com/quentez/talonjs. Otherwise if you’re using any other language, this is built off of Talon.

As a disclaimer, no quote detection algorithm is perfect, but hopefully this helps with your endeavors!

1 reply

Fronteer
July 28, 2026

Hi Rogier,
With regards to parsing out quoted content or anything similar, we’ve built a JS version of the Talon quote-detection framework that you can see here: https://github.com/quentez/talonjs. Otherwise if you’re using any other language, this is built off of Talon.

As a disclaimer, no quote detection algorithm is perfect, but hopefully this helps with your endeavors!