# Text Pipeline Playground
Pipe pasted text through a chain of transforms — grep, reject, regex replace, sort, unique, head/tail, split, join — with a tiny safe DSL. Free, private, in your browser.
## Run it
- **CLI:** `gizza tool text-pipeline-playground "Paste your text here — one record per line…" 'pipeline=grep ERROR
sort
unique'`
- **Web:** https://gizza.ai/tools/text-pipeline-playground/
- **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/text-pipeline-playground/tool.json
## Inputs
- `text` — Text _(field)_
- `pipeline` — Pipeline (one operation per line) _(field)_
- `regex_mode` — Regex mode (grep/reject patterns are regular expressions) _(field)_
- `case_insensitive` — Case-insensitive matching _(field)_
- `limit` — Max output lines _(field)_
- `on_error` — On a bad pipeline line _(field)_
## Output
- Pipeline output (text)
## Query parameters
Open the tool pre-filled and auto-run via URL:
- `text` — Text
- `pipeline` — Pipeline (one operation per line)
- `regex_mode` — Regex mode (grep/reject patterns are regular expressions)
- `case_insensitive` — Case-insensitive matching
- `limit` — Max output lines
- `on_error` — On a bad pipeline line
Example: `https://gizza.ai/tools/text-pipeline-playground/?text=Paste%20your%20text%20here%20%E2%80%94%20one%20record%20per%20line%E2%80%A6&pipeline=grep%20ERROR%0Asort%0Aunique®ex_mode=true&case_insensitive=true&limit=10000&on_error=stop`
---
## About this tool
Text Pipeline Playground lets you paste text and run it through a small chain of safe transforms: grep lines, reject noise, regex-replace fields, trim, change case, sort, dedupe, take a head/tail sample, split records, and join the result back together. It is inspired by command-line pipelines and interactive text wranglers, but it deliberately does **not** execute Python, shell, JavaScript, or user code.
Write one operation per line in the **Pipeline** box. Operations run from top to bottom, so you can build a repeatable recipe such as:
```text
grep ERROR
replace /^\S+ ERROR /!! /
sort
unique
```
For this input:
```text
2024-01-01 INFO started
2024-01-01 ERROR disk full
2024-01-02 ERROR disk full
2024-01-02 WARN low memory
2024-01-03 ERROR timeout
```
The output is:
```text
!! disk full
!! timeout
```
Supported operations:
- `grep PATTERN` keeps matching lines; `reject PATTERN` drops matching lines.
- `replace /old/new/` runs a regex replacement on each line, including `$1` backreferences. Any delimiter works, for example `replace |/|-|`.
- `prefix TEXT`, `suffix TEXT`, `lower`, `upper`, and `trim` map every line.
- `sort`, `sort -r`, `unique`, `head N`, `tail N`, and `reverse` reshape the line list.
- `split SEP` splits each line into more lines (no separator means whitespace); `join SEP` merges all lines into one line.
Turn on **Regex mode** when `grep`/`reject` should use regular expressions instead of literal substring matching. `replace` always uses regular expressions. **Case-insensitive matching** applies to grep, reject, and replace. **Max output lines** caps runaway output; **On a bad pipeline line** can stop with an error or skip malformed operations.
### Limits
The pipeline is line-oriented and text-only. It does not run arbitrary Python code, call APIs, read files, or preserve binary data. Rust regex syntax is used, so look-around and backreferences in the *pattern* are not supported, but capture groups and `$1` replacements are. Output is capped at the configured line limit after the pipeline finishes.
## FAQ
Does this run Python or shell code?
No. The name comes from Python-style text wrangling workflows, but this tool is a fixed, declarative set of operations implemented in Rust/WebAssembly. It never evaluates arbitrary code, imports packages, opens files, or starts subprocesses. That makes it safe to run entirely in the browser.
When should I turn on Regex mode?
Use Regex mode when `grep` or `reject` needs anchors, character classes, alternation, or capture-style patterns, such as `^ERROR|WARN` or `\b\d{4}-\d{2}-\d{2}\b`. Leave it off for plain substring searches — literal mode is easier for text containing punctuation like `.` or `?`. The `replace /old/new/` operation always treats `old` as a regex.
How do I remove duplicate lines?
Add `unique` after any filtering or mapping step. It keeps the first copy of each line and drops later duplicates. If you want duplicates adjacent before deduping, run `sort` first and then `unique`; if original order matters, use `unique` without sorting.
What happens if one pipeline line is wrong?
With **On a bad pipeline line = Stop**, the tool returns an error that names the failing line and explains what was expected, for example `pipeline line 2: head expected a line count`. With **Skip**, malformed lines are ignored and the remaining valid operations still run. Stop is safer for saved recipes; Skip is useful while experimenting.
Can I process CSV or JSON with this?
You can do simple line-oriented cleanup such as filtering rows, replacing delimiters, extracting a column with a regex, or turning lines into bullets. For structured CSV/JSON parsing, use a dedicated CSV or JSON tool instead — this playground treats input as plain text and does not understand quoting, nested objects, or schemas.
## Related tools
- [Find and Replace Text](https://gizza.ai/tools/find-replace/): Find and replace text with literal or regex matching, case-sensitive or not, all matches or just the first. Runs in your browser, nothing is uploaded, free.
- [Fuzzy Text Search](https://gizza.ai/tools/fuzzy-doc-search/): Paste text and fuzzy-search it — find the best-matching lines even with typos, ranked by relevance with the matches highlighted. Runs in your browser, no upload.
- [Regex Search](https://gizza.ai/tools/regex-search/): Search pasted text with a regex or literal pattern and return matching lines with line numbers, context, invert, whole-word, and highlighting.
- [Add Line Numbers](https://gizza.ai/tools/add-line-numbers/): Add line numbers to every line of text online, like nl or cat -n — custom start, step, separator, and alignment. Free and private, runs in your browser.
- [Censor Text](https://gizza.ai/tools/censor-text/): Redact a list of words (or common profanity) in any text by masking them — case-insensitive, whole-word or substring. Runs in your browser, nothing is uploaded, free.