GitHub Repository

Incremental Markdown parser that emits streams of semantic events, plus tools to manipulate them — designed for real-time rendering of streamed LLM output.

22 starsKotlin

Markanywhere – A Streaming Processor of Meanings

Name: Markanywhere – A Streaming Processor of Meanings
Availability: InStock
Author: morisil

by morisil·May 15, 2026·2 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidBig BrainNiche Gem

Event-stream parsing beats regex for streaming LLM tokens in real time.

Strengths

•Semantic event stream allows transformation before full document completion.
•Designed specifically for the latency constraints of agentic feedback loops.
•Handles mixed input formats like HTML and XML alongside Markdown.

Weaknesses

•Kotlin implementation limits adoption in the dominant JavaScript/Python LLM stack.
•Niche utility unless you are building custom streaming renderers from scratch.

Post Description

Markanywhere can parse any input, like Markdown, HTML, XML, as a stream of semantic events which can be rendered, transformed, evaluated. Works great as an interactive transport layer for the LLM inference output and agentic feedback loops.