Back to browse
GitHub Repository

Agentic Task Queue - batch process tasks with agents in parallel.

1 starsJavaScript

Agentic Task Queue – batch processing of agentic tasks

by endyai·Mar 24, 2026·2 points·2 comments

AI Analysis

●●SolidShip ItSolve My Problem

Fresh context per task prevents bloat, but agent orchestration tools already exist.

Strengths
  • Isolated contexts per task prevent accumulation and hallucination
  • Configurable concurrency pool with streaming results in input order
  • Simple stdin/stdout CLI design that pipes into existing workflows
Weaknesses
  • Anthropic-only API support limits model flexibility
  • No built-in result persistence or database output options
Target Audience

Developers running batch LLM operations on large datasets

Similar To

LangChain · CrewAI · DSPy

Post Description

https://github.com/whydna/atq

I had a huge number of database rows that each needed LLM reasoning and tool use.

Existing harnesses like Claude Code don't handle this well - give it a batch and it'll try to write you a script instead of actually reasoning over each item.

Even when it works, context bloats after each turn - gets expensive, slow, and starts hallucinating. And there's no built-in retry handling or parallelization.

So I built this tool. Define the task as a prompt, pipe in your task list, set concurrency, model, tools, etc.

Open to feedback!

Similar Projects

@push.rocks/taskbuffer – just scheduling task executions

push.rocks/taskbuffer appears to be a focused utility for queuing and scheduling short-lived tasks — useful when you need to debounce bursts or smooth calls to a rate-limited API. The npm landing (blocked by Cloudflare in my screenshot) doesn't show whether it offers persistence, retries, or clustering, so right now it looks like a handy single-process tool but not a replacement for Bull/Agenda in production workloads.

Ship ItSolve My Problem
PhilKunz
104mo ago