From interview audio to publication-ready article in minutes.
JEDI transforms raw interview recordings into structured, coherent articles — ready for editorial review. Less transcription work. More journalism.
Reclaim the hours the recorder steals.
Newsrooms today face shrinking budgets, tight deadlines and constant pressure to produce quality content. Transcription and drafting can consume 3–4 hours for every hour of recorded interview.
That’s time journalists could spend reporting, investigating, or conducting more interviews. JEDI aims to reclaim it — turning up to 60 minutes of audio into a 2,000–3,000 word article that reads like a journalist wrote it.
Three intelligent phases.
Each stage progressively refines the material — from messy raw transcript to a draft your editors can pick up anywhere.
Transcript cleanup
Merges dialogue segments, resolves split sentences, fixes grammar and typos. Every change is verified against the original transcript — no hallucinations.
Article creation
Drafts a structured piece with topic-based sections, a natural narrative flow, and follow-up questions that fill in gaps or add depth — all within the 3,000-word limit.
Refinement & verification
Checks terminology, formatting, and sentence structure, then extracts factual claims and traces them back to the source audio for built-in fact-checking.
Journalistic quality, engineered in.
It authors, not just transcribes.
Traditional AI tools stop at text conversion. JEDI understands context, structure and narrative flow — producing drafts that read like a journalist wrote them.
Every claim is traceable.
Extracted facts link back to the original recording. Accountability and fact-checking are built in, not bolted on.
Built for real newsroom workflows.
Powered by leading LLMs (OpenAI, Mistral, Claude) — ready for editors to review and collaborate.
Designed to work with journalists.
JEDI removes the drudgery. The human eye stays essential for final review and editorial judgment.
Join a select group of newsrooms shaping JEDI.
Full access during the beta, direct support from our technical team, and the chance to influence a tool built specifically for journalists.
- Full JEDI access for your newsroom during the beta
- Direct support from our technical team
- Significant time savings on transcription and drafting
- Help us optimize the model for diverse editorial contexts
A partnership for the future of newsrooms.
Innovative publishing technology for the media industry — content management, workflow intelligence and AI for publishers.
Prague-based NLP and generative-AI studio, trusted by Mediahuis, The Atlantic, RTS, The Hindu and the Czech News Agency since 2014.
Supported by FAIR — Spoke 5, High-Quality AI
JEDI is developed thanks to the FAIR EU Fund. Atex was selected among the winners of Spoke 5, focused on High-Quality AI. FAIR is promoted by the Italian Ministry of University and Research (MUR) and funded by the European Union — NextGenerationEU.
