Two-record work ledger
Maintains an append-only local ledger containing the agent's claims and observed activity, including sessions, tool calls, edits, and results.
Yoetz
Yoetz is a local-first, open-source work ledger for AI coding agents that compares agent claims with host-observed activity. It issues receipts showing what was verified, the coverage reached, and what remains open, with first-party support for Codex, Claude Code, and Cursor.
Overview
Yoetz is a local-first, open-source work ledger for AI coding agents. It records what an agent says it did alongside what the host actually observed, then checks the two records for disagreements before producing a receipt.
The receipt is deliberately bounded: it identifies what was verified, the coverage reached, and what remains open. This gives developers a more precise completion signal than an unqualified “done” claim, while keeping the service and ledger on the local machine.
Features
Completion checks
Compares published claims with recorded activity using local rules, then identifies disagreements such as evidence from a test run that predates a later edit.
Coverage-bounded receipts
Creates a receipt that records verified items, coverage, findings, and unresolved work instead of treating every completion claim as fully proven.
MCP and hook architecture
Uses MCP for agent-published information and hooks for host-observed activity, with one local service coordinating the records.
Task lineage
Supports task delegation and lineage in the terminal, including child task handles, child receipts, and one-level rollups that carry findings to the parent.
Repository coordination
Groups multiple live tasks in one repository and, with consent from both sides, informs tasks when they have declared overlapping paths.
Use cases
Reviewing agent completion claims
After an agent reports a refactor complete, Yoetz can compare the claim with edits and test evidence, flag stale runs or unverified obligations, and preserve those findings in the receipt.
Delegating repository work
A parent task can delegate work through an expiring child handle. The child publishes and checks its work, while the parent receipt rolls up child findings without treating a clean child receipt as proof of integration.
Managing concurrent tasks
When two live tasks in one repository declare overlapping paths, Yoetz can group them and, with consent from both tasks, tell each task about the other's declared overlap.
Auditing changes before acceptance
Developers can use the local ledger and receipts to inspect what an agent claimed, what the host observed, and which checks remain open before accepting a change.
Pros and Cons
Pros
- Local-first operation with an append-only ledger and no account or network required for the described workflow.
- Receipts distinguish verified coverage from unresolved findings instead of presenting a simple binary completion status.
- Open-source software licensed under Apache-2.0.
- First-party support is listed for Codex, Claude Code, and Cursor.
- Supports both individual task checks and coordination features for delegated or overlapping repository work.
Cons
- The product is alpha software and the site notes that it is new and rough in places.
- macOS is described as stable, while Linux and Windows through WSL 2 are experimental.
- The site does not publish pricing information; its pricing URL returns a not-found page.
FAQ
How is Yoetz set up?
Yoetz runs as one local service on the user's machine. An agent connects over MCP, while hooks report host activity. The site describes a recommended setup flow that copies a prompt into the agent; manual installation is also available through PyPI or npm.
What does Yoetz produce?
The agent publishes its plan, work, claims, and evidence through six MCP tools. The receipt is returned to the conversation and states what was verified, the coverage reached, and what remains open.
Which coding agents does Yoetz support?
Yoetz is intended for AI coding agents and officially supports Codex, Claude Code, and Cursor. The site also states that any MCP host can connect to the service.
Does Yoetz send project data to a remote service?
The service is local-first and the site says that nothing leaves the user's machine. A reviewer model is optional and can use the user's API key or ChatGPT plan; those runs are described as occurring from the user's environment rather than through Yoetz.
What are the current platform limitations?
The site describes Yoetz as alpha software, version 0.2.2, and says macOS is stable while Linux and Windows through WSL 2 are experimental. It also cautions that the product is new and may have rough edges.
Quick Facts
- Category
- Developer Tool
- Primary users
- AI coding agents and developers
- Deployment
- Local service on the user's machine
- Agent interface
- MCP; host activity is reported through hooks
- Official support
- Codex, Claude Code, and Cursor
- License and status
- Open source, Apache-2.0; alpha 0.2.2
Alternative a Yoetz
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
Manta AI
Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
GitBot
GitBot is a local browser-based workspace for creating reusable coding bots powered by Claude Code, Codex, or OpenCode. It helps developers run those agents across repository folders with persistent threads and configurable permissions.
hob
hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.
Ably Chat
Ably Chat is a chat API platform for building custom realtime chat applications. It supports room-based messaging, typing indicators, presence, reactions, and message updates, with usage-based pricing options for different deployment stages.