Open Source Tools for Global Communication
At Cornelius Communications, we're committed to fostering global communication and supporting language professionals worldwide. Our open source tools are designed to solve real-world challenges in translation technology, data privacy, knowledge agents, and localization workflows.
Our Commitment to the Global Language Community
We believe that technology should be accessible to all language professionals, regardless of their organization's size or budget. Our open source tools represent our commitment to:
Community Contribution
We believe in giving back to the global language community by providing free, open source tools that solve real-world problems.
Professional Quality
Our tools are built with the same attention to quality and user experience that we bring to our consulting engagements.
Industry Expertise
Each tool reflects decades of experience in translation technology, localization workflows, and global communications.
Continuous Innovation
We actively maintain and improve our tools based on user feedback and emerging industry needs.
Our Open Source Tools
Language Data Workbench (TMXmatic)
Translation TechnologyA comprehensive tool for processing TMX (Translation Memory eXchange) and XLIFF files, designed to clean and manage translation memory data efficiently. Install the app from GitHub to process files locally.
Key Features
- Duplicate management for translation memories
- TMX and XLIFF file processing
- Batch processing workflows
- Data cleaning for machine translation
- File format conversion tools
- Translation leverage operations
Use Cases
- Translation memory optimization
- Quality assurance for translation projects
- Machine translation data preparation
- Translation workflow automation
PII-Cloak
Data PrivacyWindows Desktop software for cloaking Personally Identifiable Information (PII) to protect sensitive data during translation and localization processes.
Key Features
- PII detection and masking
- Secure data handling
- Windows desktop application
- Translation workflow integration
- Privacy compliance tools
- Configurable masking patterns
Use Cases
- Secure translation of sensitive documents
- Privacy compliance in localization
- Data anonymization for testing
- Protection of personal information
SI-VidGen
Knowledge AgentsA local-first demo of how to build better Ask agents over product Help - not traditional answer bots that guess from thin retrieval. Help content is converted into OKFs (structured procedure and asset metadata), so Ask can multi-hop across grounded evidence, refuse when coverage is thin, and return summary → steps → notes with live Help links. A major enhancement over chatbot-style FAQ bots that treat documentation as undifferentiated text chunks.
Key Features
- Help corpus → OKF conversion for structured procedures
- Multi-hop Ask with refusal when evidence is weak
- Summary → steps → notes answers with Help links
- Local-first stack (Ollama, Chroma) - no required cloud AI tokens
- Create video: grounded script → Help-faithful walkthrough MP4
- Re-ingest pipeline: crawl, OKF rebuild, image library, re-index
Use Cases
- Product Help Ask agents that refuse thin answers
- Demonstrating OKF-enriched RAG over traditional FAQ bots
- Information developer and PM knowledge assistants
- Help-faithful walkthrough video generation from support issues
UI-ShotLib
Product CaptureCompanion project to SI-VidGen that walks a web-based product and captures its screens into a reusable image library. Useful on its own for UI catalogs, QA, localization visual checks, and bot-building - and as the screenshot source SI-VidGen binds into product faithful walkthrough videos.
Key Features
- Automated walk of web-based product UIs
- Screenshot capture for new and changed screens
- Reusable image library for catalogs and bots
- Companion asset pipeline for SI-VidGen video composition
- Supports QA and localization visual review workflows
Use Cases
- UI screen catalogs for product documentation
- QA and localization visual regression support
- Screenshot libraries for Help-faithful video agents
- Training and bot-building over real product surfaces
AgentProtect
AI SecurityOrchestrates layered security models for AI agents - firewall checks for prompt injection, jailbreak, and credential leakage; social-engineering and reputation protection; data and privacy controls; output sanitization; and infrastructure safeguards such as I/O validation and rate limiting.
Key Features
- AI firewall: prompt injection, jailbreak, credentials
- Social engineering and reputation protection
- Malicious code, privacy, and fraud checks
- Output/error sanitization and log redaction
- Memory sanitization and credential monitoring
- I/O validation and rate limiting
Use Cases
- Hardening production Ask and agent deployments
- Reducing prompt-injection and data-exfiltration risk
- Sanitizing agent outputs before they reach users or logs
- Layered guardrails across agent platforms
Graph Help
Help SystemsTurns classic help into a learning experience - concept graphs and interactive exploration so people (and agents) can navigate product knowledge by relationships, not only keyword search.
Key Features
- Concept-graph navigation of help topics
- Interactive learning-oriented help experience
- Relationship-first exploration beyond keyword search
- Prototype suitable for product and documentation teams
Use Cases
- Modernizing classic help into guided learning paths
- Exploring how concept graphs improve discovery
- Prototyping next-generation documentation UX
- Complementing Ask agents with structured topic maps
Explore All Our Open Source Projects
Visit our GitHub profile to discover more open source tools and contributions to the global language technology community.
View GitHub ProfileNeed Custom Tools for Your Organization?
Our open source tools demonstrate our technical capabilities. We can build custom solutions tailored to your specific global communications and localization needs.
Discuss Custom Solutions