open-test-data

Synthetic people for use as test fixtures: names, short physical descriptions, and AI-generated profile portraits. Every record is fabricated. Use it wherever you need realistic-looking sample users without touching real personal data.

No real people. Names are randomly generated. Faces are produced by an image model from a text description and do not depict any actual person. Any resemblance to a real individual is coincidental.

Layout

manifest.json                       index of every person in the set
person/
  <name_hash>-<sequence>/
    person.json                     the record
    person_log.json                 image generation history
    profile/
      base.webp                     neutral reference portrait
      professional.webp             studio portrait derived from base

Each person lives in one directory named by its key. The key is a 16-hex-character hash of the name followed by a sequence number, for example 00f3f6c55434db3e-1. The sequence lets two distinct people share a name.

manifest.json

The entry point. Lists every person with the fields most consumers need.

{
  "format_version": 1,
  "entries": [
    {
      "key": "00f3f6c55434db3e-1",
      "display_name": "André Vaillant",
      "images": {
        "base": "profile/base.webp",
        "professional": "profile/professional.webp"
      }
    }
  ]
}

Image paths are relative to the person's directory, so the full path is person/<key>/<path>.

person.json

Field	Type	Meaning
format_version	int	Schema version, currently 1
key	string	<name_hash>-<sequence>, matches the directory name
name_hash	string	16 hex characters derived from the name
sequence	int	Disambiguator for people sharing a name hash
first_name, last_name	string	ASCII or romanized form
display_name	string	Preferred rendering, may use native script or accents
names	object	Per-locale spellings keyed by locale, e.g. ja_JP, fr_FR. Empty when the ASCII form is the only one
gender	string	M or F
locale	string	Locale the name was drawn from, e.g. en_US, ja_JP
look	string	Coarse appearance category used to steer image generation
appearance	string	One-sentence physical description used as the image prompt
appearance_seed	int	Seed that produced the appearance text
images.<role>.path	string	Relative path to the image
images.<role>.image_seed	int	Seed used to generate that image
images.<role>.reference	string	Role whose image was used as a likeness reference

Example:

{
  "format_version": 1,
  "key": "00f3f6c55434db3e-1",
  "name_hash": "00f3f6c55434db3e",
  "sequence": 1,
  "first_name": "Andre",
  "last_name": "Vaillant",
  "display_name": "André Vaillant",
  "names": { "fr_FR": { "first_name": "André", "last_name": "Vaillant" } },
  "gender": "M",
  "locale": "fr_FR",
  "look": "white",
  "appearance": "white man in his 50s, light skin, buzzed silver hair, grey eyes, no glasses, friendly grin, a trimmed goatee",
  "appearance_seed": 2528958472,
  "images": {
    "base": { "path": "profile/base.webp", "image_seed": 3019527744 },
    "professional": { "path": "profile/professional.webp", "image_seed": 312866366, "reference": "base" }
  }
}

person_log.json

An append-only history of image generation for the person. Each entry records the role, the action (generate or regenerate), the seed, the full prompt, the model, a timestamp, and the reference role if one was used. The current image for a role is the result of the latest entry for that role.

{
  "format_version": 1,
  "key": "00f3f6c55434db3e-1",
  "entries": [
    {
      "ts": "2026-09-17T00:03:12Z",
      "role": "base",
      "action": "generate",
      "image_seed": 3019527744,
      "prompt": "Neutral reference portrait, head and shoulders, ...",
      "model": "gemini-2.5-flash-image",
      "model_version": "gemini-2.5-flash-image"
    }
  ]
}

Images

Two portraits per person, both 384×384 WebP with no embedded metadata.

• base: head and shoulders, facing the camera, plain light grey background, even lighting, natural expression. This is the canonical likeness.
• professional: a studio portrait of the same person with a colored backdrop and business or smart-casual clothing, generated using base as a reference image.

What's in the set

People	59
Images	118
Locales	en_US, fr_FR, ja_JP, es_MX, en_IN, tr_TR, zh_CN
Gender	37 M, 22 F
Image model	gemini-2.5-flash-image

Usage

import json
from pathlib import Path

root = Path("open-test-data")
manifest = json.loads((root / "manifest.json").read_text())

for entry in manifest["entries"]:
    person_dir = root / "person" / entry["key"]
    person = json.loads((person_dir / "person.json").read_text())
    portrait = person_dir / entry["images"]["professional"]
    print(person["display_name"], person["locale"], portrait)

Notes

• All content is generated. Names come from a locale-aware random name generator; descriptions and portraits come from seeded generation and can be reproduced from the recorded seeds and prompts.
• look and appearance exist to make the image set visually diverse and to keep each person's portraits consistent with each other. They are generation inputs, not demographic data about anyone.

License

Licensed under the Apache License, Version 2.0. See LICENSE for the full text and NOTICE for the attribution notice.

Copyright 2026 AppCove, Inc.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

This covers the whole repository, the JSON records and the images alike. If you redistribute any part of it, including a single portrait, include a copy of the license and retain the notice.