Releases·Tool Release
Show HN: AgentRun: DSL to turn agents into Workflows
A Parcha engineer launched AgentRun on Hacker News, an open-source DSL powering their grep.ai harness that converts repeatable parts of agent jobs into inspectable workflows. Steps can mix tool calls, code, LLM-driven routing decisions, and full agents only where investigation is needed, cutting the cost of running a complete agent loop. It ships with a scripted demo, GitHub examples, a Pi extension, and a background video.
8 engagement1 sourceWed, Sep 23, 2026, 07:42 PM
AI Overview
- Parcha AI open-sources AgentRun, the DSL its grep.ai harness uses to turn repeatable agent work into workflows mixing tool calls, code, and agents
- The harness reads traces and retro notes agents leave to decide which steps can become workflows, avoiding full agent loops where unneeded
- Example research workflow splits questions into subquestions, researches them in parallel, screens evidence with Jev, then has an agent write the report
- Runs as a Pi extension to build, inspect, and run workflows; scripted demo requires no API keys (github.com/Parcha-ai/agentrun)
Related
Products·building
Stageflow — configurable pipeline for multi-stage LLM agent workflows
Stageflow is a pipeline tool that breaks complex agent jobs into stages, each with a fresh agent session doing one pre-configured task. LLMs power each stage, with clean handoffs keeping behavior scoped. It targets developers who want reliable agent workflows without drift, plus optional human approval gates.
Stageflow is a pipeline tool that breaks complex agent jobs into stages, each with a fresh agent session doing one pre-configured task. LLMs power each stage, with clean handoffs keeping behavior scoped. It targets developers who want reliable agent workflows without drift, plus optional human approval gates.
6 engagement1 sourceredditTue, Sep 8, 2026, 01:20 PM
Releases·Tool Release
Hanesu: an experimental workflow layer for AI coding agents released on GitHub
A developer released Hanesu, an experimental workflow layer for AI coding agents, on GitHub. The tool aims to address agent context loss and task repetition by providing structured workflows and persistent task state. It is designed for features, bug fixes, and patches, with error logging to aid iterative improvement.
A developer released Hanesu, an experimental workflow layer for AI coding agents, on GitHub. The tool aims to address agent context loss and task repetition by providing structured workflows and persistent task state. It is designed for features, bug fixes, and patches, with error logging to aid iterative improvement.
1 engagement1 sourcehackernewsThu, Jul 23, 2026, 06:58 AM
Releases·Tool Release
Developer releases Stageflow, a configurable multi-stage pipeline for checking AI vibecoding slop
Developer releases Stageflow, a multi-stage pipeline for keeping AI coding output in check
Developer releases Stageflow, a multi-stage pipeline for keeping AI coding output in check
1 engagement1 sourceredditTue, Sep 8, 2026, 02:18 PM
Products·building
Visual multi-agent workflow editor exporting runnable CrewAI code
AgentGraph Studio is a visual editor for designing multi-agent workflows and exporting them as runnable Python code. It uses LLMs to power agents like researchers and analysts, and solves the problem of making complex agent orchestration accessible and verifiable for developers.
AgentGraph Studio is a visual editor for designing multi-agent workflows and exporting them as runnable Python code. It uses LLMs to power agents like researchers and analysts, and solves the problem of making complex agent orchestration accessible and verifiable for developers.
3 engagement1 sourceredditThu, Aug 20, 2026, 10:42 AM
Products·launched
Agent Process: browser extension that turns website workflows into WebMCP tools
An open-source browser extension that records a website workflow as a JSON script and exposes it as a WebMCP tool for AI assistants. Instead of an agent re-deriving clicks via UI inspection each time, users teach the process once and call it with parameters (e.g. create_lead in a CRM). It works on any website without native MCP support or an API, and much of the extension was built with Codex.
An open-source browser extension that records a website workflow as a JSON script and exposes it as a WebMCP tool for AI assistants. Instead of an agent re-deriving clicks via UI inspection each time, users teach the process once and call it with parameters (e.g. create_lead in a CRM). It works on any website without native MCP support or an API, and much of the extension was built with Codex.
1 engagement1 sourceredditFri, Sep 4, 2026, 07:00 PM