Honest math:
The "Published Benchmark" mode uses our FastAPI retrieval test — 214,312 tokens vs ~480 tokens, which works out to roughly $1.08 saved per benchmark-style lookup at Claude Sonnet pricing. Real sessions vary wildly (most queries don't dump 200K tokens). The "Conservative" mode uses the per-iteration delta from our Vue+Firebase A/B test instead, which is closer to typical day-to-day work. Pick the mode that matches your skepticism level. Either way, payback lands in literally days of normal use.
Verified by practitioners
Four signals. Four buyer emotions.
Efficiency
5×
"Roughly 5× more efficient context retrieval."
Artur Skowroński · VirtusLab
Reasoning Quality
Tokens for thinking, not retrieval
"Preserves your context budget for actual reasoning."
Sion Williams
Structural Depth
Queries native tools can't answer
"Structural questions you simply can't ask Grep or Glob."
Traci Lim · Amazon Web Services
Scarcity Economics
Context is the scarce resource
"The whole game is what you choose not to put in the prompt."