The desk in one screen
Capital markets executives stop stitching three tools and a spreadsheet. P&L, pullthrough, exceptions, and a daily summary are co-located by construction.
Live P&L roll-ups, daily sparklines, pullthrough scoring, lock-outcome materialization, anomaly callouts, and an AI daily briefing — for executives who need the desk in one screen and analysts who need to drill into any number.
Overview
The Hedge Cockpit is a server-rendered capital-markets dashboard backed by the hedge-service, lock-service, and ai-service. P&L aggregates over the materialized lock-outcome view; daily sparklines hydrate from the same store; pullthrough scoring runs on every lock with an explainable rationale; anomalies bubble up through the cockpit and via SSE; and the AI daily briefing summarizes what moved with cited links into the source events. Every number is one click away from its source trace, and the URL state is shareable for desk handoffs.
Aggregate market value, delta-since-open, exposure by program / vintage / investor, all materialized from the lock-outcome store.
Per-segment daily series for trend visibility without leaving the cockpit. Hover any point to drill into the day's events.
Predicted pullthrough on every active lock with an explainable scorer — see exactly which inputs drove the score.
A morning summary describing what moved and why, with citations into the source events. Generated by ai-service, audit-chained, and rate-limit-capped.
Drift, confidence drops, missing fills — each surfaces in the cockpit and as an Exception Inbox item with a suggested next action.
Filter set, time range, drill-down path — all in the URL. A desk handoff is a link, not a screen share or a 'walk me through it' meeting.
How it works
Numbered steps from input to output. Each step maps to a specific subsystem you can inspect via OpenTelemetry.
locks.created / locks.funded / locks.cancelled, pricing.computed, and ratesheets.activated land in NATS JetStream with correlationId and arrive at the hedge-service materializer.
The materializer maintains a fact table of lock outcomes (funded, fallout, extended) with attribution by program, vintage, investor, branch, and LO.
On lock create + on ratesheet activation, the scorer updates predicted pullthrough on active locks with feature attribution.
Page-level RSC composes P&L, sparklines, exceptions, and the AI briefing in one render. SSE keeps the page live; no client-side polling.
ai-service builds the briefing from the lock-outcome view and the day's anomalies, cites source events, and writes the response to audit.
Click any number to land on the underlying event trace. Share the URL with a colleague; export to PDF for the morning meeting.
Hands on
Live cURL sample — copy, paste, ship.
# Pull today's P&L roll-up via API — same data the cockpit renders
curl 'https://api.ratestack.com/v1/hedge/pnl/today?desk=primary' \
-H "X-API-Key: $RATESTACK_KEY"
# Returns:
# {
# "asOf": "2026-05-26T13:42:11Z",
# "deskId": "primary",
# "headline": { "marketValue": 1.23e8, "deltaSinceOpen": -42100, ... },
# "byProgram": [ { "program": "CONFORMING_30YR", "mv": ..., "delta": ... }, ... ],
# "byVintage": [ ... ],
# "anomalies": [
# { "severity": "warn", "lockId": "lck_8a7c",
# "message": "Pullthrough confidence fell 0.18 vs prior session",
# "suggestion": "Re-run pullthrough score after the 14:00 sheet" }
# ],
# "briefing": "The desk is down 42k since open, driven by a 4 bps drift on 7yr ..."
# }Why this matters
Capital markets executives stop stitching three tools and a spreadsheet. P&L, pullthrough, exceptions, and a daily summary are co-located by construction.
Anomalies push to the cockpit via SSE — no polling, no missed minute. The Exception Inbox is one click away with AI-suggested resolutions.
Every number drills into a source event with correlationId; the AI briefing's citations are first-class, not decorative.
Frequently asked
Sub-second from the originating event to the cockpit. SSE delivers the push; the materializer keeps the lock-outcome view warm; the cockpit re-renders the affected segments only.
Yes. PDF export is built-in; a Slack delivery hook is available on Business and Enterprise tiers.
The briefing is grounded on the day's lock-outcome view, anomaly stream, and ratesheet activations — only data within your tenant. Every claim cites a correlationId you can verify in the audit chain.
It is read-only by design — no state mutations originate in the cockpit. If hedge-service is unreachable, the page degrades to the last cached snapshot with a 'stale as of …' indicator.
Yes, with capability scoping. Cross-desk visibility is gated by the capability catalog; a desk-A user with cross-desk read can land on the desk-B cockpit URL and see exactly desk B's view.
How we compare
Specific angles, not generic feature checklists. Each row links to a longer side-by-side; we're transparent about where competitors are the better choice.
Hedge cockpit is not a separate product — it's the same audit chain and event stream that drives pricing and locks.
See the side-by-sideCockpit ships shareable URL state and an AI daily briefing grounded on your tenant data, not a static report PDF.
See the side-by-sideComparisons reflect each vendor's public positioning. Where a fact is unverifiable, we mark it "Depends" or "Unknown" instead of guessing.
Related capabilities
Ready to see it on your data?
We'll spin you a sandbox, load your actual ratesheets, and walk you through this capability against your top scenarios.