Brett Terpstra launches Marked Share Markdown publishing service

// 1h agoPRODUCT LAUNCH

Brett Terpstra launches Marked Share Markdown publishing service

Marked Share is a free publishing and collaboration service designed to share Markdown and TextBundle documents with rendered styling, attachments, and short links via markedb.in. Originally conceived alongside the Marked 3 Mac preview utility, it functions as a standalone platform accessible via a browser editor, an open REST API, or desktop app integrations. In addition to private drafts, password protection, and custom slugs, Marked Share includes collaborative inline highlighting and commenting for asynchronous document reviews, as well as collection feeds with native RSS, Atom, and Micro.blog syndication.

// ANALYSIS

While traditional pastebins offer zero collaboration and CMS platforms demand excessive setup, Marked Share carves out a compelling niche by turning plain text into an interactive, shareable review medium without lock-in. Native support for TextBundle and image attachments removes the biggest historic limitation of sharing raw Markdown files, while inline highlighting and reader notes enable Google Docs-style editorial feedback directly on rendered web documents. Standalone design and API access make it easy to pipe documents from Obsidian, Neovim, VS Code, or command-line scripts, and public collections paired with RSS, Atom, and Micro.blog support double as an ultra-lightweight personal blogging engine.

// TAGS

markdownpublishingtextbundlewriting-toolsdevtoolcollaborationblogging

DISCOVERED

1h ago

2026-09-12

PUBLISHED

6h ago

2026-09-12

RELEVANCE

6/ 10

AUTHOR

Brett Terpstra

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE51m ago

Mercury Agent gates permissions across 40+ tools

Mercury Agent announced an update emphasizing its suite of over 40 built-in tools paired with a granular, tiered permission architecture. Designed to balance autonomous agency with developer safety, Mercury allows safe read-only operations and inspection commands to proceed without interruption, while enforcing explicit confirmation gates for sensitive actions such as file modifications, arbitrary shell executions, directory scoping, and external network calls.

MODEL1h ago

OpenAI GPT-Live commoditizes front-end voice orchestration

Satvik Bansal analyzes OpenAI's GPT-Live in the five-layer voice AI stack, explaining how its native turn-taking and interruption handling act as an ultra-responsive host while commoditizing layer-4 orchestration like LiveKit and Pipecat. Regional and domain-specific voice AI startups remain resilient, however, as enterprise moats shift to deep workflow integrations, data residency, and low-bandwidth telephony economics.

NEWS1h ago

Meta Muse reveals six subagent worker types

A peek under the hood of Meta's newly launched personal AI assistant Muse reveals its internal multi-worker agent architecture. When queried inside the mobile application, Muse clarified that rather than maintaining fixed persona rosters, it spawns ephemeral task-specific subagents across six specialized archetypes: General Worker, Browser Operator, Deep-Research Agent, Scheduled Worker, Artifact Builder, and Workflow Orchestrator.