Rectifa Labs

Understanding
: Information
at its source

Rectifa develops proprietary tools to analyze information dynamics on digital channels: detecting weak signals, verifying sources, and modeling the spread of narratives.

Status
Prototype phase
Active Development · 2026
First tool
Recheck
URL, Text, and Image Analysis · FR / EN
Field
Information Risk Detection
Social Media · Digital Channels

R&D rooted in real-world applications

Methods of manipulating information are evolving faster than the tools used to detect them. Our approach: build testable prototypes, measure their effectiveness in real-world scenarios, and then iterate.

01

Field observation

Analysis of how information spreads on social media to identify the most active sources of misinformation.

02

Rapid prototyping

Every hypothesis is translated into a testable tool. The prototype is the real testing ground, not the report.

03

Evaluation and Iteration

Our tools are evaluated based on criteria such as accuracy, bias, and real-world usage. The results guide the development of each new version.

Prototype · Active

Recheck: Analyze information in seconds

Rectifa’s first tool. It assesses the level of information risk associated with content (URLs, text, or images) shared on social media, in both French and English.

🔗
URL AnalysisContent posted on YouTube, TikTok, and other social media platforms.
📄
Text AnalysisClassification of an excerpt and detection of misleading signals.
🖼️
Image AnalysisIdentification of manipulations and misleading visual contexts.
⚠️
Risk ScoreGraded result: Low / Moderate / High, based on fact-checking sources.
Access the prototype
recticheck.netlify.app
🔗 URL
📄 Text
🖼️ Image
Optimized for social media · FR / EN
e.g., youtube.com/watch?v=… or tiktok.com/@account
Assess a level of risk
Low
Moderate
High
Working prototype

This tool is currently undergoing active research. The results are preliminary and are subject to ongoing evaluation.

The problems our R&D team is working to solve

01
🕸️

Network detection

Modeling the dynamics of disinformation spread on platforms that distribute news content.

Propagation · Graphs
02
🤖

AI-powered multimodal analysis

Approaches capable of analyzing text, images, and video together.

NLP · Computer Vision · Large Language Models
03
📊

Risk classification

Qualitative scores that reflect the inherent uncertainty of any analysis of information.

Scoring · Evaluation
04
🌍

Multilingualism

Incorporating linguistic and cultural considerations from the design phase. FR/EN, with plans to expand to African languages.

FR · EN · Multilingual
05
🔍

Primary sources

Institutional fact-checking frameworks to ground analysis in verifiable data.

Fact-check · Sources
06
🧭

Tool design

UX research for interfaces that are accessible to everyone, from experts to the general public.

UX · Popularization
2025
Research Scope
Defining priorities, reviewing existing solutions, and identifying gaps.
Early 2026
Development of Recheck v0
First prototype for analyzing URLs, text, and images with risk scoring.
2026 · In progress
Evaluation and Iteration
User testing, model improvement, multilingual expansion.
2027
Embeddable version
APIs and modules for media outlets, institutions, and partner platforms.

Research open to partnerships

We collaborate with academic institutions, media outlets, and civil society organizations that share our goals.

If you are a journalist, researcher, or organization interested in testing our prototypes, please contact us.

Contact Us
Rectifa Labs

The news deserves better

We are building the tools of tomorrow to make information more reliable, more transparent, and easier to understand.