What we deliver.
Finished data inside your tools. Priced per task, per batch or per project.
AI answer rating
Two AI answers in, a choice and a reason out, against your guidelines.
Per rated pair
Coding evaluation
AI-written code reviewed, tests written and solutions rated. Python and JavaScript.
Per reviewed task
Answer writing
The answer a model should give, written and checked against your guide.
Per written answer
Safety testing
Attempts to make a model fail in Vietnamese, each written up with its severity.
Per finding or per batch
Speech data
Speech recorded and transcribed in Northern, Central and Southern accents.
Per recorded hour of audio
Test translation
English AI tests turned into natural Vietnamese and checked for cultural fit.
Per test item
Fact-checking
Each claim in an AI answer checked against a source and marked right or wrong.
Per checked answer
The cards show examples of each kind of work, not client data.