Controls you can evidence.

Numbers you can defend in a review, with provenance, replay, and published invariants.

Your problem

The parts you keep working around.

01

A control you cannot evidence is a finding, and most data vendors offer no correctness guarantee at all.

02

Reproducing a historical figure months later requires the context that produced it, not just the answer.

03

Halts, pauses and session boundaries are exactly the moments risk cares about, and exactly what generic feeds drop.

The proof

Nine invariants, asserted against production.

The suite is open source and runnable by anyone, against us or against any competitor implementing the same endpoints. Results are published continuously.

Liquidated notional, hourly

StreamData · stream_type: EVENTS

Inside the 08:11 window

type: liquidation · block order

TimeVictimCoinSizeMark px
08:11:42.1080x3fa8…c41dBTCliquidated 1.8420066,412.0
08:11:42.9510x91d2…77beETHliquidated 41.06003,188.4
08:11:44.3100xb05c…2e9aBTCliquidated 0.9715066,371.0
08:12:01.0770x44e1…9c08HYPEliquidated 8,420.038.115
08:12:03.5620x7d93…f1a6ETHliquidated 12.33003,184.1

Illustrative events in the live-feed format. Each liquidation carries the victim wallet, liquidated size, price, and mark price, in block order. Funding payments arrive on the same stream.

Integration sketch

Reproducible months later.

Every record carries the code version that produced it, and reprocessing a range at that version is byte-identical. Daily merkle roots let you prove nothing was altered after the fact.

Read the docs
# The invariant suite is public and runs against production —# ours, and any competitor implementing the same endpoints. npx @robinhoodrpc/invariants \  --endpoint https://api.robinhoodrpc.io/v1 \  --sample 500 # ✓ balance_ui == balance_raw x multiplier_at(block)# ✓ historical queries never use the current multiplier# ✓ session boundaries match the published calendar# ✓ every SEQUENCED record reaches a terminal state# ✓ premium_bps recomputes from its component fields# ✓ tape_sequence is strictly monotonic, no gaps
Where to start

The plan that fits.

Scale

$200/moRecommended

gRPC streaming, full history, and every sink.

Signed snapshots and deterministic replay are Scale features. Audit export and SOC 2 artefacts come with Enterprise.

Compare both plans

Bring us into your review.

We can walk through provenance, replay and snapshot verification with your risk team before anything goes into production.