Upload prompt files, paste prompt text, or drop a JSON config. We'll figure out the rest.
Drop files here or click to browse
.txt, .md, .json, .yaml — prompt files or agent configs
We identified these agents and inferred how they connect. Adjust anything that's off.
Analyzing your agents...
We generated 10 synthetic test cases from your agent prompts. Remove any that don't fit.
Generating test cases...
Spectral will auto-generate evaluation dimensions and start scanning. Here's the summary.