Building intelligent publishing tools at the intersection of editorial workflows, AI-driven insights, and personalized reader engagement. Thoxt is designed to empower writers to create, collaborate, and for readers to discover high-quality content at scale.
Thoxt translates editorial expertise into a high-scale, open platform for user-generated content (UGC). Writers deserve tools that understand what makes content impactful, not just clicks, so we built editorial intelligence workflows from scratch.
We aim to bridge the gap between traditional editorial judgment and modern digital publishing, enabling writers to craft high-quality, engaging articles with automation tools and optimized workflows.
The Thoxt editor is a low-latency, cloud-native editor optimized for real-time authoring and sub-second rendering. It enables co-author support, multimedia embedding, versioning, and autosave without page reloads.
Thoxt Editor is a creative partner, not just a text box. Built on Quill, it extends standard editors with advanced features designed for professional publishing:
Compared to Notion, Medium, or Substack, Thoxt provides a fully integrated publishing studio whith SEO and metadata automation, collaboration tools, and multimedia capabilities. Compared to publishing on traditional blogs, e.g. Wordpress, Thoxt can save upto 5 hours per post.
Editor workflow: Draft → Insights → Short video preview (optional) → Co-author/editor edits (optional) → Publish → Feed → Reader engagement & Analytics → Feedback loop.
Most social media or content recommendation engines surface a combination of recent and engagement-producing posts. Thoxt’s recommendation engine aligned algorithmic output with human storytelling intuition. It is built on built three layers of intelligence:
Recommentations Workflow: Separate news vs non-news → Order news by narrative time → Keep semantic similarity as the backbone → Avoid burying context-setting pieces → Avoid old news overpowering new developments.
Future roadmap: it is the responsibility of a news platform to keep users informed on what's generally important in the world right now. The best version of Thoxt recommendation engine will keep users informed, help discover new topics and form new hobbies/skills, present opposing views, not just to keep users in their own filter bubble.
Writers often struggle to find relevant, timely topics aligned with their audience. Thoxt’s topic suggestion engine solves this by:
Ideate Workflow: External feed → Clustering → Classification → Writer history → Suggestion ranking.
Future roadmap: a 'novelty-aware ideation engine' that predicts emerging reader interests while staying true to a writer’s voice and audience.
Thoxt integrates a journalism-aware feed from around the web:
Our hybrid editorial feedback system combines AI, NLP libraries, and in-house rule-based algorithms to provide real-time insights that mimic an editor’s review. We measure:
Unlike Grammarly or Hemingway, Thoxt evaluates journalistic nuance for search-engine and social media discoverabilty.
Thoxt combines multiple technical building blocks to power editorial intelligence:
We intentionally design journalism-aware recommendation systems, where loss functions prioritize editorial quality and reader engagement over pure click metrics.
Thoxt combines aggregation, custom ML, and rule-based heuristics to mirror human editorial judgment, creating scalable tools for high-quality, user-generated journalism.