Workflow

Pipeline

Chain decode, format, convert and clean steps, see every intermediate result and share the recipe, never the input. Runs in your browser.

Your input stays in your browser. The chain runs in a background worker on your device. Nothing is uploaded, stored or put in the URL, and nothing you paste is ever run as code. A recipe link contains only the steps and their options, never your input.

Steps

No steps yet. Add one below, load a preset or paste a recipe.

    Ready. Paste some input, then add steps.

    Result

    Recipes

    A recipe is the list of steps and their options as JSON. Copy it, paste one here and press Apply, or share a link. The link and the recipe never contain your input or any result.

    What it does

    Pipeline joins existing DevToolEra transformations into one chain. Add steps, reorder them, switch them on or off and set their options. Every step shows its own output, so you can see where a chain changes the text. If a step fails, the chain stops at that step and tells you which one and why. The result is updated as you type.

    Each step uses the same function as the matching tool (JSON Formatter, JSON Repair, YAML ⇄ JSON, TOML ⇄ JSON, JSON ⇄ TOON, JSONPath Tester, Sort Lines, Slug Generator, String Escape, Hex Viewer, JWT Decoder, Hidden Unicode Detector, Secret Scanner, AI Output Sanity Check, AI Code Risk Linter). Base64, URL encoding, HTML escaping, text to bytes, hashing and timestamps use the same browser functions as their tools, because those tools keep their logic inside the page script.

    What it checks

    • Limits: input and every result are limited to 5,242,880 characters (5 MiB of text), and a chain can have 20 steps.
    • Recipes: a pasted recipe or a link is checked strictly: only known step types, known options and values of the right type and range are accepted. A recipe is data and is never run as code.
    • Suggestions: the tool looks at the result and offers next steps when it looks like JSON, a JWT, Base64, hex, percent-encoding, a PEM block, a Unix timestamp or a UUID, or when it contains hidden characters or something that looks like a key. These are guesses and are labelled heuristic.

    What it cannot catch

    • A chain only transforms text. It does not check that the result is correct, and “Valid JSON” means the text parses, not that it has the shape you need.
    • JWT steps only decode. They do not verify a signature, so the content is not proven genuine.
    • Some steps are guesses by nature: JSON repair and the LLM clean-up can add or drop data, and a step says so in its notes.
    • Suggestions are guesses from the shape of the text. A plain number can look like a timestamp, and a hex string can also be Base64.
    • HTML unescape reads numeric references and a short list of named ones; the HTML Entity tool reads all named entities.
    • The Code risk step makes a report from pattern checks. It does not run or understand the code, and no result means no pattern matched, not that the code is safe.

    Limits

    Steps run in a background worker, so a large input does not freeze the page. A chain that takes longer than about a minute plus a little time for each 50,000 characters is stopped. Outputs larger than 300,000 characters are shown only in part; the copy buttons copy everything. The recipe link is limited to about 16,000 characters.

    Privacy

    This tool runs entirely in your browser. What you enter is processed on your device and is not sent to a server. The page does not write your input to local storage, cookies or the URL. A recipe link holds only the steps and their options, and the page does not send your input anywhere.

    Share this tool

    Your data is not included in this link.