Why run them in one go?

Calling the tools one at a time means a person or an AI copies the intermediate values across. A 64-character hash where one character goes wrong changes everything downstream — and it is invisible to the eye. Chaining them keeps the intermediate values from ever leaving.

Which tools can I put in?

Any tool whose core is separated out: hashes, Base64, VAT, business days, age, Korean keyboard fix, unit conversion and more. Give a name that does not exist and it tells you before running.

If something goes wrong, do I find out where?

Yes. Each step reports its own result, and on failure it says which step it stopped at. Showing only the final value would leave nothing to trace back through.

Does it work from an AI agent?

chain_run on the MCP server (karmolab-mcp) does the same job. It uses the same code as the screen, so the results cannot diverge.

Chain tools

Name: 도구 묶어 쓰기
Author: KarmoLab

Each result feeds the next · every step visible · up to 8 steps

도구 묶어 쓰기

도구 여러 개를 이어서 한 번에. 앞 결과가 다음 도구의 입력이 됩니다

단계 (JSON)

"$1" = 1번 결과 통째로 · "a{{1}}b" = 글 안에 끼우기 · 최대 8단계

예시가 들어 있습니다 — 실행을 눌러 보세요

Runs several tools one after another. Each step’s result becomes the next step’s input, so there is no copying intermediate values by hand.

How to use it

Write the steps as a JSON array (tool · op · args)
Use "$1" for the whole previous result, or "a{{1}}b" to slot it inside text
Hit run and each step reports its result on its own line
It works from a URL too: /karmolab/t/chain/?op=run&steps=…

Common questions

Why run them in one go?: Calling the tools one at a time means a person or an AI copies the intermediate values across. A 64-character hash where one character goes wrong changes everything downstream — and it is invisible to the eye. Chaining them keeps the intermediate values from ever leaving.
Which tools can I put in?: Any tool whose core is separated out: hashes, Base64, VAT, business days, age, Korean keyboard fix, unit conversion and more. Give a name that does not exist and it tells you before running.
If something goes wrong, do I find out where?: Yes. Each step reports its own result, and on failure it says which step it stopped at. Showing only the final value would leave nothing to trace back through.
Does it work from an AI agent?: chain_run on the MCP server (karmolab-mcp) does the same job. It uses the same code as the screen, so the results cannot diverge.

Other tools

Stays on your device What you enter is handled inside your browser only — nothing is stored or sent anywhere.

This tool page states where what you enter goes — see the badge above. All tools · KarmoLab · Who made this · Built with AI. View source