Back to Changelog
Improvement

In-App Document Chunking

Moved our document chunking pipeline from an external service into the core Rails app. Faster processing, fewer dependencies, simpler architecture.

Previously, uploaded documents were sent to a separate microservice for chunking and indexing. We moved this logic into the main Rails application, eliminating an external dependency and simplifying the entire pipeline.

Faster processing

Documents now process inline without the network overhead of calling an external service. Most files are chunked and indexed within seconds of upload.

Simpler operations

One fewer service to deploy, monitor, and scale. The chunking logic lives alongside the rest of the application code, making it easier to debug issues and ship improvements.

Same quality

The chunking algorithm itself is unchanged — documents are split at natural boundaries (headings, paragraphs) to preserve context. The only difference is where the code runs.

Make customer support your competitive advantage

Start using SupportUnicorn today — free plan available.