effect-agent
Yielded Agent: durable Effect agent threads in TypeScript, listed for a TypeSafe decision provider for typed question sets.
Overview
Yielded Agent (effect-agent) builds durable agent threads in TypeScript with Effect and Effect AI. You define schemas, instructions and tools; the runtime runs the model and tool loop, validates output and can retain accepted work across process restarts. Threads, runs and attempts are distinct, and model context, history and recovery state have separate lifetimes, so long conversations can be compacted without discarding committed history. The application supplies authorisation, tool handlers, model services and storage.
According to the list summary, it includes a TypeSafe decision provider for typed question sets and optional model selection. The README excerpt reviewed does not describe this provider, so how Jev is used inside the runtime is not documented here. The package is in beta. Policies bound each run by maximum turns, tool calls and duration.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Typed question sets answered by the TypeSafe provider (e.g. which model to use)(inferred) | — |
How it works
- Define an agent with schemas, instructions, tools and policy
- Provide model and history layers (optionally the TypeSafe provider)
- Runtime runs the bounded model/tool loop
- Durable receipts allow recovery after restarts
Strengths
- Durable, recoverable agent execution
- Bounded work policies and approvals
- Native Effect AI integration
Caveats
- TypeSafe provider not described in the README
- Beta package
Getting started
bun add @yielded/agent@beta effect, then follow the getting-started guide at yielded.dev.At a glance
- Type
- Project
- Status
- beta
- Language
- TypeScript
- Runtime
- Effect library (Bun/Node)
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
An Effect Agent TypeSafe decision provider for typed question sets and optional model selection.
logicrw/awesome-jev-projects
Frequently asked questions
What is Yielded Agent?
A durable agent runtime for Effect that survives process restarts.
Where does Jev fit?
Per the listing, as a TypeSafe decision provider for typed questions and model selection.
Is it stable?
It is published as a beta.
Related projects
LangChain's Python framework, which includes an optional TypeSafeClassifier runnable plus model-routing and risky-tool middleware using Jev.
PluginproductionPythonLangChain.js, which includes an optional TypeSafeClassifier runnable and middleware for bounded routing and tool-call checks with Jev.
PluginproductionTypeScriptVercel's filesystem-first agent framework, whose experimental evaluate and autoModel paths default to Jev via AI Gateway.
ServicebetaTypeScriptComfyUI custom node that asks Jev noul, choice or score questions inside image-generation workflows, e.g. to pick a workflow template.
ProjectprototypePythonCommunity .NET 10 SDK giving one API for choice, score and assessment over Microsoft-Decision-1, Ollama or native Jev.
PluginbetaC#LangChain blog guide on wiring Jev into an agent harness as a dangerous-action classifier and model router.
Articlediscussion