Trigger
Start a run from a manual input, webhook, or supported market signal.
Signal · intelligence · control
Build autonomous market workflows that watch your inputs, call the decision engine, enforce your risk policy, and record every outcome.
The loop
One workflow carries context from the first trigger to a result you can inspect. You design the graph; Peregrine runs the same guarded path every time.
Start a run from a manual input, webhook, or supported market signal.
Bring market data and portfolio context into one structured decision input.
Call the selected model and validate its output against a strict trade schema.
Apply venue rules, position limits, global policy, and the kill switch.
Keep the trigger, reasoning, status, and confirmed execution result together.
Two starting points · one loop
Use the same visual builder with a different starting point. The contracts, risk checks, and activity trail stay consistent.
Respond to what you hold
Start with your current exposure. When a signal changes the case for a market, the workflow adds context and proposes a controlled response.
Position response
Illustrative run · no live order
Transparency by default
Peregrine keeps the decision contract, policy checks, and runtime state visible. When execution is not confirmed, the interface says so.
0
Visible stages
From incoming trigger to a recorded result
0
Global policy
One source of truth for limits and venue access
0%
Schema checked
Model decisions must match the shared trade contract
See the trigger payload, structured decision, status, and any confirmed execution result in Activity.
Position caps, venue access, and the global kill switch stay visible in one risk policy.
Missing model access falls back to hold, while API and venue failures remain explicit instead of becoming fake orders.
Your next signal is coming
Connect a trigger, choose the decision path, set your boundaries, and see the complete run in one workspace.