DeniableUse cases
Test the cases your systems still have to handle.
Different workflows fail in different places. Deniable groups pre-generated synthetic documents around the testing job you want to inspect, so you can start with a concrete case instead of a generic file.
Choose a testing job
Start with the failure you want to inspect.
General testing
Edge case testing
Probe the awkward inputs that routine examples leave out. Designed scenarios make missing context, conflicting details, and unexpected structure easier to inspect.
Explore edge-case testing →Retrieval and grounding
RAG evaluation
Evaluate how a document-grounded workflow finds and uses source material. Keep the relationship between a query, a document, and an answer visible in your own test process.
Explore RAG evaluation →Use cases
Document categories for the workflows you need to test.
Support Tickets
Customer issues, follow-ups, and inconsistent context for support workflows.
Explore category →Product Reviews
Mixed sentiment, incomplete details, and uneven product language.
Explore category →Social & Chat Logs
Short turns, slang, context shifts, and conversational noise.
Explore category →Meetings & Call Transcripts
Turn-taking, corrections, omissions, and spoken context.
Explore category →Legal Documents
Clauses, references, and details that depend on surrounding language.
Explore category →HR Documents & Resumes
Profile information, timelines, and varied document structure.
Explore category →Finance & Invoices
Totals, line items, dates, and structured fields.
Explore category →Code Q&A
Technical questions, snippets, and the context around an answer.
Explore category →Long-form Documents & Ebooks
Sections, references, and broader document context.
Explore category →Mixed Edge Cases
Cross-format inputs, conflicting context, and workflow seams.
Explore category →Testing & Evaluation
A separate path for evaluation work.
RAG evaluation and edge case testing belong together when the question is how a system behaves under difficult conditions. Benchmarking and CI/CD testing can join this category when their datasets and workflows are ready.
Applications
Use the data for the workflow you are building.
The categories describe what a document is. These application pages describe what teams can test with it.
Machine Learning & NLP
Training and evaluation for sentiment, intent, entities, and text classification.
Explore application →Chatbot & Conversational AI
Multi-turn context, dialogue quality, and safe response workflows.
Explore application →RAG & LLM Evaluation
Retrieval, grounding, citations, and long-context behaviour.
Explore application →Email & Anti-spam
EML parsing, threading, headers, encodings, and message classification.
Explore application →CRM & Support Workflows
Routing, automation, dashboards, and customer-service operations.
Explore application →Text Analytics & BI
Aggregation, topics, sentiment trends, and reporting pipelines.
Explore application →Privacy & Redaction
Sensitive-field detection, masking, redaction, and document handling.
Explore application →Load & Performance
Repeatable document volumes for ingestion, indexing, and API stress runs.
Explore application →Education & Research
Practical synthetic material for NLP, data science, and testing exercises.
Explore application →Marketing & Social Listening
Controlled reviews and social-style text for analysis workflows.
Explore application →