Inteply

Inteply is an AI research and publishing platform that helps journalists and newsrooms go from a topic to a source-backed, publication-ready article in minutes, without sacrificing accuracy or editorial control.
The Market Problem
Journalists spend hours manually researching, validating sources, and fact-checking before writing. Existing AI tools offer “prompt in, article out” shortcuts that newsrooms can’t trust. Inteply solves this with a credible, structured workflow built around how editors actually work.
From Experimental AI to Venture-Ready Product
What started as an AI generation experiment was rapidly iterated into a full multi-step editorial pipeline. Each sprint iteration added a new layer of production readiness — source validation, factual filtering, multi-model orchestration, job tracking, and analytics — transforming a prototype into a commercially deployable newsroom tool.
Autonomous AI Pipeline
The core product is a multi-agent workflow where different models handle different jobs:

• Live source discovery and ranking
• AI-powered source validation, clustering, and deduplication
• Factual filtering to reject low-value sources before generation
• Fact extraction per source, deduplication across sources, and coverage sufficiency checks
• Final source-backed article generation only after all evidence is assembled
  Model size is matched to task complexity — smaller, faster models for structured extraction and filtering tasks, larger models for final generation — optimizing for speed, cost, and quality across the pipeline.
  Technologies & Frameworks
  Frontend: Next.js, React, TypeScript, Tailwind CSS, shadcn/ui
  Backend: Node.js, TypeScript, Express.js
  Database & Auth: Supabase
  AI models based on task: Claude, Mistral, Gemini
  Observability: Langfuse — used to trace the full LLM pipeline across all models, capturing prompt input, model output, token usage, cost, and latency end-to-end
  Search: Brave Search API
  Infrastructure: Caching, retry logic, rate limiting, job tracking, and a custom AI call monitoring layer
  Scalability
  Multi-model routing means costs scale efficiently with volume. Graceful fallbacks handle blocked sources. Job tracking and monitoring support production-level reliability at scale.
  Traction
  Inteply is live and in the go-to-market phase. We are actively reaching out to journalists and newsrooms, using MuleRun to automate outreach workflows.