← back to ecosystem
app beta version 1.0.0-beta.1 hex unreleased package jidoka elixir ~> 1.18

Jidoka

Developer-friendly LLM agent harness built on Jido and Jido AI

Beta support. Public iteration with faster feedback loops before final release shape settles.

GitHub ★4 Issues
View package metadata source →

RELATED PACKAGES

Builds on
Ash Jido

Jidoka builds directly on this ecosystem package.

ecosystem page github
Jido

Jidoka builds directly on this ecosystem package.

ecosystem page hexdocs
Jido AI

Jidoka builds directly on this ecosystem package.

ecosystem page hexdocs
Jido Browser

Jidoka builds directly on this ecosystem package.

ecosystem page hexdocs
Jido Character

Jidoka builds directly on this ecosystem package.

ecosystem page github
Jido Eval

Jidoka builds directly on this ecosystem package.

ecosystem page github
Jido MCP

Jidoka builds directly on this ecosystem package.

ecosystem page github
Jido Memory

Jidoka builds directly on this ecosystem package.

ecosystem page github
Jido Runic

Jidoka builds directly on this ecosystem package.

ecosystem page hexdocs

AT A GLANCE

Developer-friendly LLM agent harness built on Jido and Jido AI
Agent and workflow DSL surfaces for higher-level runtime composition
Guardrail modules for input, output, and tool boundaries
Trace and AgentView concepts for observing agent execution

DEEP DIVE

Overview

Jidoka is an experimental agent harness and DSL layer built on Jido and Jido AI. It explores higher-level composition patterns for agent views, guardrails, and workflow runtime behavior.

Purpose

Jidoka is for proving out developer-friendly agent-building patterns above the lower-level Jido runtime and package stack.

Boundary Lines

  • Owns its DSL, guardrail, AgentView, and harness-level runtime composition.
  • Depends on core Jido, AI, memory, browser, MCP, and workflow packages.
  • Does not define the canonical low-level Jido runtime contracts.

Major Components

Agent DSL

Provides higher-level authoring surfaces for defining agent behavior.

Guardrails

Includes input, output, and tool guardrail modules for runtime boundaries.

AgentView And Tracing

Explores observability and capture patterns for understanding agent execution.