Product design | UX/UI
The Václav Project is an AI-powered web service that supports newsgathering workflows
Radio Free Europe/Radio Liberty
The problem
Journalists spend too much time using multiple tools to transcribe, translate, and summarize content.

The solution
A single easy-to-use tool for solving many tasks related to newsgathering: transcription, translation, summarization, and video downloading.
Challenges
Ensuring accuracy across languages, managing AI costs, and maintaining reliable access to video content.
My role
Led UX/UI design, including user interviews, needs analysis, user personas, feature prioritization, user flows, prototyping, usability testing, and UI kit development. Later transitioned into project management, focusing on usage and cost estimation, stakeholder communication, and user onboarding.
Desktop version
Product Capabilities
One unified workspace for multilingual content processing: from transcription and translation to summaries, subtitles, and tags.
  • Transcription
    Convert audio and video into structured, timestamped text with speaker identification.
  • Subtitles
    Generate multilingual subtitles with adjustable formatting and export options.
  • Translation
    Translate content across multiple languages with linked original and translated text blocks.
  • AI Tags
    Automatically generate contextual tags in multiple languages.
  • Summarization
    Generate concise AI-powered summaries to quickly understand long-form content.
  • Task Management
    Track processing progress, receive notifications, and access recent results.
Discovery & Research
One unified workspace for multilingual content processing: from transcription and translation to summaries, subtitles, and tags.
  • Competitive analysis
    Market landscape, existing tools, competitor features and UX best practices.
  • User Research
    Interviews, needs, pain points, workflows and hypothesis validation.
  • AI Model Evaluation
    Language tests, multilingual accuracy, processing speed and model capabilities.
  • Research Synthesis
    Key insights, validated hypothesis, design requirements and product opportunities.
Competitive analysis
Happy Scribe
DeepL
GoodTape
Turboscribe
Discovery & Research
User Personas
& Workflows
  • Conducted user interviews to understand journalists' needs, pain points, and existing workflows.
  • Identified distinct user groups based on their goals and behaviors, and mapped user flows to reflect different use cases.
  • Validated initial product assumptions through user interviews and identified opportunities for improvement.
  • Synthesized research findings into a prioritized needs matrix to guide feature development and product decisions.
Discovery & Research
Core Product Flow
A headline's purpose is to quickly and briefly draw attention to the story. It is generally written by a copy editor, but may also be written by the writer, the page layout designer, or other editors.
PERSONA WORKFLOW
Journalist
Turn long interviews into searchable, editable transcripts and extract relevant quotes for articles.
Transcription, timestamps, speaker identification, synchronized playback, search, and export.
Key features
PERSONA WORKFLOW
Multimedia Editor
Prepare video content for editing, voice-over, localization, and subtitling.
Transcription, subtitles, synchronized timestamps, subtitle settings, and export.
Key features
PERSONA WORKFLOW
Standards Editor
Review source material and translated content for accuracy, context, and editorial compliance.
Linked transcripts and translations, synchronized navigation, and source comparison.
Key features
PERSONA WORKFLOW
Localization Editor
Quickly assess whether regional news is relevant to an English-speaking audience before committing to translation and adaptation.
Summarization, multilingual translation, linked content blocks, and export.
Key features
From user needs to product decisions
We synthesized interviews and product feedback into actionable insights, role-based requirements and research-informed development priorities.
  • “I need to work with long interview recordings, not only links.”
    Journalist, focus group
  • “A short summary helps me decide whether a regional story is worth translating.”
    Central Newsroom Journalist, focus group
  • “I need timecodes synchronized with text and subtitles in .srt format.”
    Video production, focus group
Key research insights
Recurring patterns in newsroom workflows revealed what mattered most to different user groups.
  • Speed & efficiency
    Faster processing of lengthy interviews and multilingual stories.
  • Rapid evaluation
    Summaries help editors decide if a story deserves full translation.
  • Content navigation
    Timecodes, speaker labels and structured transcripts support editing.
  • Language coverage
    Regional newsrooms required reliable support for less common languages.
User needs matrix
A compact cross-persona view of recurring needs. Dots represent illustrative role relevance, not research frequency or numerical survey results.
User need
Journalist
Multimedia
Standards
Localization
Transcription
Translation
Summaries
Subtitles
Speaker identification
File upload
Export options
Parallel comparison
UX Design & Prototyping
01
Content Input & Language Selection
We develop the real-time scene, implement interactions, and iterate through tests. We focus on creating solutions that bring real value from the very beginning
02
Language Selection
We develop the real-time scene, implement interactions, and iterate through tests. We focus on creating solutions that bring real value from the very beginning
02
Transcript & Translation Workspace
We develop the real-time scene, implement interactions, and iterate through tests. We focus on creating solutions that bring real value from the very beginning
03
AI-Powered Features
Summaries, subtitles, and contextual tags within one unified workspace.
Interaction &
Usability Testing
Tested key user interactions and workflows with target users, using feedback to improve navigation, feature discoverability, and overall usability.
UI Kit &
Component  library
Created reusable UI components and consistent interaction patterns for multilingual content processing, ensuring a cohesive experience across transcription, translation, and other AI-powered workflows.
~400
Users with SSO access
~750
Person-hours saved monthly (estimated)
4–5×
Faster content processing (estimated)