LEDGIT

Identity-linked audit trails for every AI agent · 3 CLI commands

Without vs With LEDGIT

	Without	With LEDGIT
Human oversight	None or weak prompts	Hardware-signed approval
Audit trail	Basic logs, easy to fake	Immutable on Hedera HCS
Proof	“The agent said so”	Provable human signature
Regulatory	Not viable	Enterprise-ready
Accountability	Blame the AI	Chain of custody
Safety	Rogue action risk	Controlled & traceable
Agent logging	Prompts & outputs	Hardware-proofed action trail

Agents without LEDGIT run on trust. Agents with LEDGIT run on proof. Existing logging platforms track prompts or logs — LEDGIT provides cryptographically verifiable human proof at ~$1/mo for 10k actions.

Install LEDGIT

Prerequisites

Node.js 20+ or Bun

Install

$ git clone https://github.com/cbonoz/ledgit.git

$ cd ledgit && bun install

$ npm link

$ ledgit --version

$ ledgit setup

Interactive setup: connects to Hedera, creates your first agent and HCS topic, generates action config. Bring your own ENS name or use a local identifier — no subname service needed.

Configure your agent

$ ledgit tools schema

Outputs an OpenAI/Claude-compatible tool definition. Pipe this into your agent framework to give it access to propose actions that require human approval.

Powered By

LEDGIT·GitHub·FAQ·Architecture·ETHGlobal New York 2026