Skip to content
Outrich Data

Chat dataset format converter and validator

Paste or drop a fine-tuning dataset and convert it between OpenAI chat JSONL, ShareGPT, Alpaca, prompt and completion pairs, and conversation JSONL. Every record is checked for missing roles, empty content, wrong turn order and length, with problems listed by line number. Valid records and rejects download as separate files. It runs in your browser.

How to use it

  1. 01Paste JSONL or a JSON array, or drop a file. The source format is detected per record, or pick it yourself.
  2. 02Choose the target format and the token limit for a single record.
  3. 03Read the problems table. Each rejected record is listed with its line number and the reasons.
  4. 04Download the valid records in the target format, and the rejects file to fix or discard.

What it does not do

  • Token counts are an estimate at about four characters per token. Use your model's tokenizer for exact limits.
  • Only system, user and assistant roles pass validation. Records with tool calls or image content are rejected.
  • Alpaca and prompt and completion formats hold one exchange. Multi-turn records are rejected when you convert to them.
  • Slack and email threads have no roles. The first speaker becomes the user and everyone else the assistant.

Questions

About the chat dataset format converter.

Is my data uploaded anywhere?
No. The tool runs entirely in your browser. Your text and files are never sent to Outrich Data or anyone else, and nothing is stored after you close the tab. You can confirm it by turning off your internet connection after the page loads: the tool keeps working.
What makes a record invalid?
A missing or unknown role, empty content, a system message that is not first, a first turn from the assistant, two turns in a row from the same role, a record that ends on a user turn, or a record over your token limit.
What is in the rejects file?
One JSON object per rejected record with its line number, the list of problems and the original record, so you can repair it and run it again.
Can it read the output of the other parsers here?
Yes. Conversation JSONL from the ticket parser maps customer turns to user and agent turns to assistant. Thread JSONL from the Slack and email parsers is also accepted.
Is it really free?
Yes. There is no account, no limit on the number of uses and no watermark. Files up to 50 MB are supported.

More free tools

Personal data scrubber

Strip names, emails, phones, cards, keys and addresses from text and files.

Dataset profiler

Rows, columns, tokens, date span, duplicates and personal-data columns in one report.

Column hasher and remover

Drop sensitive columns or replace them with salted hashes that still join.

CSV, JSON and JSONL converter

Convert between CSV, TSV, JSON and JSONL without breaking quoted cells.

JSONL validator

Find the broken lines in a JSONL file and see which keys are missing.

Duplicate remover

Remove repeated rows, records or log lines, keeping the first of each.

Token counter

Size text the way AI labs do: tokens, words, characters and lines.

Data card generator

Write the one-page description every data buyer asks for.

Data rights pre-check

A tailored checklist of the contracts and questions to clear before a sale.

Slack export parser

Turn a Slack workspace export into clean thread JSONL with pseudonymized users.

Email thread parser

Convert .mbox and .eml files into threaded JSONL without quoted replies.

Ticket export parser

Map help desk exports to customer and agent turns in conversation JSONL.

Log parser

Turn web server, syslog, logfmt and JSON logs into structured JSONL or CSV.

HTML and document text extractor

Extract clean text from HTML, Markdown, XML and JSON files in batches.

Re-identification risk checker

Measure k-anonymity and l-diversity, then preview fixes that lower the risk.

Secrets scanner

Find API keys, tokens, passwords and private keys in a repo or export before you share it.

Delivery manifest builder

Create a SHA-256 checksum manifest for a dataset delivery, with one hash to compare.

Delivery verifier

Check received files against a manifest: match, changed, missing or extra.

File checksum calculator

SHA-256 of any file, with a compare box that turns green or red.

Train, validation and test splitter

Split a dataset into train, validation and test files without leaking related rows across splits.

Dataset sampler

Pull a random, stratified or first-N sample from a file, including files too large to open.

Schema inferrer and data dictionary

Infer field types and stats from a file and export JSON Schema, a data dictionary and SQL.

Text chunker

Split documents into overlapping chunks for retrieval or training, with heading context.

Dataset diff

Compare two versions of a dataset: added, removed and changed records, plus schema changes.

Encoding and mojibake fixer

Repair garbled characters, convert files to UTF-8 and strip invisible characters.

Column toolkit

Rename, split, merge, clean, filter and sort columns with a reusable pipeline of steps.

Cleaned it up? Sell it to Outrich.

Outrich Data buys company data directly. Check in five minutes whether yours qualifies, with no call.

esc