Natural-language search

What it is

A search bar that takes plain-English queries and translates them into the same filter shape the screener uses — available from every page. The translation runs server-side; results come back as a filtered screener URL you can refine further. Keyboard shortcut: ⌘K on Mac, Ctrl+K on Windows.

Queries that work well

The translator understands bond-domain concepts: currency, sector, rating tier, country, maturity range, coupon range, yield range, issuer name, structure (144A / Reg S). Concrete examples:

• investment-grade USD utilities maturing 2030-2035
• BBB EM sovereign in EUR yielding above 5%
• Brazilian corporates with coupon between 6 and 8
• Apple Inc bonds — issuer name lookup
• Japanese banks 5-10 years
• Short-duration high yield in dollars
• Healthcare bonds with yield over 7

Queries that don't translate

The translator is a filter generator, not a research assistant. These don't work (yet):

• Holdings questions: which funds hold AAPL bonds (use Sherpa Signal on the bond detail page).
• Math / analytics: what's the duration of treasury 10-year (use the bond detail page).
• Forecasts / opinions: which bonds will outperform — no model output here.
• Free-text issuer descriptions: bonds from companies making semiconductors — sector taxonomy is fixed, not vibes-based.

Quotas

If the result misses

Natural-language search is a faster entry point for common shapes. If it doesn't return what you expected, the filter rail is always available, always exact, and never gated. Refine any filter directly to narrow further.