# Batch Executor: Repetitive Jobs Done Item by Item > The Batch Executor is the AI batch processing helper in the Ai1 platform by MyZone AI: when one of your other agents needs the same simple step run across many items, such as resizing photos or checking links, it runs each item and hands back a plain results table, making no decisions and contacting no one. Canonical page: https://myzone.ai/pages/agents/ai-batch-worker-agent Part of Ai1, by MyZone AI. Book a Batch Executor walkthrough: https://calendly.com/d/ct6h-tcy-8qf/ai1-demo?a1=Batch%20Executor&utm_source=myzone.ai&utm_medium=agent-page&utm_content=ai-batch-worker-agent-final Last updated: 2026-10-05. Reviewed by the MyZone AI team. Repetitive jobs done item by item, with a status for every result. ## At a glance - **Who uses it:** Your other Ai1 agents; you keep asking the agent you already use - **Typical jobs:** Resizing or converting images, link and file checks, permission sweeps and checks another agent names - **What comes back:** A one-line summary, a results table, counts and the files it wrote - **Judgement calls:** Sent straight back to the agent that asked - **Scope:** Only the files a job names; an unusually large batch stops for a report first - **What it never does:** Contact a person, commit, push or deploy code, or install software ## What the Batch Executor does - **Handles bulk image jobs:** Resizes photos and changes file formats across a whole set, then lists the new sizes and flags any file that failed or could not be opened. - **Checks links and files:** Works through a list of links or files and records a status for each one, with broken or missing items shown first. - **Runs checks another agent names:** When one of your agents needs the same check run over many files, it runs exactly that check on each file and records the result. - **Reports evidence, not opinions:** Each batch comes back with a one-line summary, a results table, counts of passes and failures, and a list of the files it wrote. - **Handles downloaded content carefully:** Anything it downloads or opens is logged by status, size and fingerprint. It ignores instructions hidden in that content and flags anything that looks suspicious. - **Sends judgement calls back:** If a job needs a decision, or is not really a batch, it goes straight back to the agent that asked for it. ## How the Batch Executor differs from your main AI agent Your main agent could resize photos or check links, but it would spend its time and running cost on repetitive steps instead of your task. The Batch Executor takes that work off its hands: it runs exactly the step it is given on each item, touches only the files the job names and returns a status for every item. It makes no decisions, never contacts a person and never ships anything. ## How it works 1. **An agent hands over a batch** (When an agent has repetitive work): One of your agents passes it a list of items and the exact step to run on each one. 2. **Batch or bounce** (Before any item runs): It checks the job really is a repeatable step. Anything that needs judgement is returned to the sender straight away. 3. **Run every item** (Item by item): It runs the step on each item in turn and records the outcome, touching only the files the job names. 4. **Report back** (As soon as the batch ends): The agent that sent the batch receives the summary, the results table, the counts and the list of files written. ## When to use it - **Your website agent needs a set of product photos ready for the web:** It resizes and converts the whole set, then lists each new file along with any photo that was too small or damaged. - **Every link on a site needs checking before a relaunch:** It runs bulk link checks and returns a status for each link, with the broken ones at the top for a person to fix. - **An agent asks for something that needs a design decision:** It does not guess. The job goes back to the sending agent with a single line explaining that a decision is needed. - **The same check has to run across many files:** It applies the check to each file and reports which passed and which failed, a simple form of repetitive tasks automation. ## What you get - A one-line summary of the batch - A results table with a status for every item - Counts of what worked and what failed - A list of the files it wrote - Notes on anything suspicious in downloaded content ## Example: product photo and link batch Example with a fictional company. Names, people and figures are invented to show the agent's output. Any resemblance to a real company or person is unintended. What it was asked: Resize our new product photos for the web and check every shop link on the relaunch list, then tell us what failed (asked of the shop's website agent, which handed the batch to Batch Executor). Batch Executor ran one real batch on synthetic files: it measured, fingerprinted (SHA-256) and resized 23 generated product photos with standard image tools, and checked 24 links once each against a disposable local copy of a shop, logging every exit code. The faults were seeded on purpose for the example; nothing was published, deleted or changed on any live site, and the report went back to the website agent that sent the batch; every fix is left to a person on the team. Run date: 30 September 2026. ### What it found - 5 of 24 shop links return "page not found": the jugs collection and four product pages. A person on your team fixes or removes each one before relaunch. - 3 of 23 product photos are narrower than the 1,200 pixel minimum in the brief, and 1 file is damaged and cannot be opened, so those 4 need a fresh photo. - 17 photos became web-ready files, cutting their full-size weight by 92% (11.8 MB to 0.95 MB). 2 other photos were exact copies stored under a second name. ## Guardrails - It makes no decisions, never contacts a person and never ships anything. A person on your team makes every change that follows from its report. - It never commits, pushes or deploys code, installs software or changes server settings. - It only touches the files a job names. If a batch is unusually large or would reach other files, it stops and reports back first. ## Frequently asked questions ### How do bulk link checks work before a website relaunch? Every link on the list is checked and given a status, with broken or missing ones shown first. In Ai1 by MyZone AI, your website or development agent hands the list to the Batch Executor, which returns a results table so a person on your team can fix each broken link. You do not work with it directly. It takes work from your other agents rather than from people: you keep asking the agent you already use, and that agent hands over batch jobs when it makes sense. ### Which repetitive tasks suit automation by an AI agent? Simple steps that repeat across many items and need no judgement: resizing or converting images, checking links and files, permission sweeps and running the same check across many files. Anything that needs a decision is not a batch, so the Batch Executor sends it back to the agent that asked. ### Why hand batch jobs to a separate AI worker? Simple, repeated steps run in a small separate worker instead of taking up the time of your larger agents. That keeps the agent you work with focused on your task and helps keep running costs down, and every item still comes back with its own status, so nothing is quietly skipped. If a batch is too big or reaches beyond its files, it stops and reports back before going further. It only works on the files a job names, and a missing tool is reported so the System Administrator can install it. ### Is the Batch Executor available today, and will it contact my team? Yes. It runs on your Ai1 server as a behind-the-scenes helper, running the same simple step across many items for your other agents and handing back a plain results table. It never contacts a person. If something needs your input, it returns the job to the agent that sent it, and that agent asks you. ## More on this topic - [Multi-Agent Development: Why the Future of Software Is Built by AI Teams](https://myzone.ai/pages/blog/multi-agent-development): Discover how multi-agent AI systems are transforming software development. Teams of specialised AI agents build products 10x faster at lower cost. ## About Ai1 Ai1 is the AI operations platform by MyZone AI, where each client runs on its own private server. The Batch Executor is included on every Ai1 level, including Developer Core, with no per-agent charge. It works alongside the other Ai1 agents on your account. Pricing: https://myzone.ai/pages/services/ai1-pricing. Security: https://myzone.ai/pages/security.