Skip to main content
Precipiq is the financial accountability layer for production AI. Every decision your agents make — a discount offered, a refund approved, an email sent — is recorded on an append-only, hash-chained ledger and linked to its downstream dollar consequence: revenue earned, cost incurred, or liability opened. The result is a forensically auditable trail from “the model said X” to “the customer was charged Y.”

Quick Start

Install the SDK, log your first decision, and see it on the dashboard in five minutes.

Core Concepts

Understand Decision Records, Financial Events, Consequence Links, and the Hash Chain.

REST API Reference

Full endpoint reference with interactive try-it forms and multi-language code samples.

Integrations

Connect Stripe, QuickBooks, LangChain, and CrewAI to your Precipiq ledger.

​
How Precipiq works

1

Log a decision

When your AI agent picks an action, the SDK writes a Decision Record — inputs, outputs, confidence score, and a human-in-loop flag — to the append-only ledger.
2

Chain the record

Each record hashes its predecessor’s hash into its own prev_hash field, creating a tamper-evident chain. A single verify_chain call detects any alteration instantly.
3

Ingest financial events

Stripe and QuickBooks webhooks automatically create Financial Events when payments clear, refunds fire, or disputes open. You can also POST events directly from your own billing system.
4

Link and measure

The attribution engine connects decisions to events. Your AI P&L dashboard shows live revenue, cost, and open liability attributed per agent — and threshold alerts fire when exposure spikes.

​
Why teams choose Precipiq

Forensic auditability

Export a cryptographically signed evidence bundle for regulators, auditors, or post-incident review — without depending on production logs.

AI P&L visibility

Answer “is this AI making or losing money?” with a live dashboard that attributes dollars to specific agents and action types.

Compliance-ready records

Decision records map directly onto EU AI Act Article 12 fields and support U.S. discovery with per-decision queryability and hash-chain integrity proofs.

Drop-in integration

Wrap a function with @pq.track or attach a LangChain callback — the SDK handles batching, retries, and fallback logging without touching your transport layer.