PipeLedger AI

DELIVERY

Three delivery surfaces. One security layer.

Serve governed financial data to any AI agent, app, or pipeline — through MCP, REST, or the CLI — all enforced by the same policy.

The problem

Every team wants GL data somewhere different — an assistant in Claude, a dashboard over REST, a job in CI. Wire each up separately and security drifts surface to surface.

How it works

Published once, read by every surface

Your ERP / GL
Raw ledger data from any source system
PipeLedger transformations
one policy · access · masking · approvals
Published dataset
snapshot · As of 2026-Q2 · certified
MCP, REST, and CLI all read the published dataset
MCP
For AI assistants

Ask questions in plain language and get governed answers — e.g. in Claude.

REST
For apps & dashboards

Feed the tools your team already uses every day, over a versioned API.

CLI
For autonomous agents

High-frequency infrastructure that runs unattended in any pipeline.

Source provenance
run_id · published-at · certified-by — on every response
Drill-down handle
/v1/drill/{result_id} — signed
Protocol
MCP spec 2025-11-25 conformant
Surfaces
MCP, REST & CLI — identical scope and masking