AppsSynthetic Data

Where data lands · Autessa Synthetic Data

JSONL and CSV dataset files

Autessa Synthetic Data produces downloadable JSONL and CSV dataset files. Choose the "Dataset file" target, pick a format, and when the job finishes "Export dataset" gives you a temporary download link to the assembled file.

A core synthetic data capability, done properly
Start free
edge-case-invoices.csvSAMPLE
invoice_no,customer,amount,currency,status
INV-20418,Fenwick Supply,1290.50,USD,Paid
INV-20419,"Ortega, Lane & Co",0.00,USD,Void
INV-20420,Kestrel Labs,-45.00,EUR,Credit
INV-20421,Blue Fern Cafe,18990.00,USD,Overdue
CompletedFile · CSVSAMPLE

Generated 1,000 rows as a CSV dataset.

⇩ Export dataset

Opens a temporary download link to the assembled file.

The problem

Test data often lives outside the database

Fixtures for a test suite, inputs for a data pipeline and training sets for a model all need files, not table rows. Copying rows out of a database to get a file adds a step and a chance to get the format wrong.

What you get

  • JSONL (the default) or CSV.
  • Up to 5,000 rows per job.
  • CSV columns follow your schema order, and values are formatted the same way a table insert would format them.
  • The file is assembled when the job finishes and downloaded through a temporary link.
  • Dataset files are the default target on the CLI and API, written directly unless you ask for review.
  • Review first works for files too: exclude rows before the file is assembled.

How it works

How does jSONL and CSV dataset files work?

  1. Choose Dataset file

    Pick the target and choose JSONL or CSV.

  2. Define the columns

    Propose a schema, apply the training pairs preset, or add columns yourself.

  3. Generate

    Start the job. Batches are generated in parallel and staged.

  4. Export

    When the job completes, press "Export dataset" to open the download link.

Every option

Dataset file options

OptionWhat it does
File formatJSONL or CSV. JSONL is the default.
File nameThe dataset name.
ColumnsRequired. From the schema editor, a proposal or the training pairs preset.
DownloadA temporary download link, available once the file has been assembled.
CLI / API defaultOmitting the target creates a JSONL dataset file written directly.

Example

Create and download a dataset file from the CLI

Replace the ids with your own.

A JSONL dataset from the CLIEXAMPLE
# 1. Draft columns from a description (creates nothing)
autessa synthetic-data propose-schema \
  --description "Customers of a project-management SaaS, with realistic signup notes"

# 2. Start a job: a JSONL dataset file by default, staged for review here
autessa synthetic-data create \
  --name "Demo customers" \
  --description "Customers of a project-management SaaS, with realistic signup notes" \
  --rowCount 500 \
  --writeMode REVIEW \
  --columnDefinitions '[{"columnName":"company","columnType":"SHORT_TEXT","required":true},
    {"columnName":"plan","columnType":"SINGLE_SELECT","allowedValues":["Starter","Team","Business"]},
    {"columnName":"signup_note","columnType":"LONG_TEXT"}]'

# 3. Poll until AWAITING_REVIEW, then review, approve and download
autessa synthetic-data get --jobId <jobId>
autessa synthetic-data preview --jobId <jobId> --batchIndex 0
autessa synthetic-data approve --jobId <jobId>
autessa synthetic-data export --jobId <jobId>

Permissions

Who can do what?

WhoCan
People in your companyCreate datasets, see the company list of jobs, open any job and review, approve, append, cancel or discard it.
API keysCan run quick generate, propose schemas and run dataset file jobs. Table targets need a signed-in person.

Connected

How jSONL and CSV dataset files connects to the rest of Autessa

SYNTHETIC DATA FEATURE

LLM training data

The most common use for JSONL output.

SYNTHETIC DATA FEATURE

REST API

Generate fixture files from CI with an API key.

SYNTHETIC DATA FEATURE

Quick generate

For 50 rows or fewer, download CSV or JSONL instantly.

FAQ

Questions about jSONL and CSV dataset files

How long does the download link last?

It is temporary. Press "Export dataset", or run `autessa synthetic-data export --jobId <id>`, whenever you need a fresh link.

How are multi select values written in CSV?

The same way a table insert stores them, so the file matches what you would see in a table. JSONL keeps the generated values as JSON.

Do dataset files count against database storage?

Every job uses AI usage for generation. Rows written into tables also count against your database storage limit, and if that limit is reached the job ends partially completed.

Try jSONL and CSV dataset files in Autessa Synthetic Data

Autessa Synthetic Data produces downloadable JSONL or CSV dataset files of up to 5,000 rows, delivered through a temporary download link.

Start free

Describe it · review it · write it · or download JSONL and CSV

100% customizable, code and all

Make Autessa Synthetic Data fully your own.

Use it out of the box, then change anything. Ask Autobot, Autessa’s AI builder, for a new field, screen, report, workflow or agent and it builds it into your copy of the app. Or open the code and edit it yourself. Nothing is locked: the app, its data and its code are yours.

Start free and make it yours