3 Releases 3 Tags

RSS Feed

• v1.2.0 8d88342fc5

  Compare

  v1.2.0 Stable

  naomi released this 2026-04-28 14:58:08 -07:00 | 0 commits to main since this release

  Overview

  Minor update refining the character reference sheet and image generation prompt for more accurate results.

  Changes

  Reference Sheet

  • Updated to the latest version of Naomi's reference sheet

  Character Prompt

  • Removed the barefoot instruction — footwear is now determined by the scene like clothing and accessories
  • Clarified that vampire fangs are only visible when the mouth is open

  Notes

  Requires a Google Gemini API key. You will be prompted to enter one on first launch.

  Downloads

  • tatsumi_1.2.0_x64-setup.exe
    20 MiB

    2026-04-28 14:57:57 -07:00
  • Source Code (ZIP)
  • Source Code (TAR.GZ)
• v1.1.0 5c65eb9ae5

  Compare

  v1.1.0 Stable

  naomi released this 2026-04-13 21:07:11 -07:00 | 2 commits to main since this release

  Overview

  Quality-of-life update adding image input support across all modes, user-selectable aspect ratio and resolution, and generated thread names.

  Changes

  Image Input

  • Art and Avatar modes now accept an optional image alongside the text prompt — paste or upload a reference and Gemini will factor it in
  • Replace mode now has an optional notes field on the initial prompt, so you can include extra context with your source image

  Thread Configuration

  • Aspect ratio — Art mode lets you pick your aspect ratio (1:1, 4:3, 3:4, 16:9, 9:16, 21:9) when starting a new thread
  • Resolution — All three modes let you choose output resolution (1K, 2K, 4K) when starting a new thread

  Thread Names

  • Threads are now named with randomly generated adjective-noun combos (e.g. Crimson Reverie, Neon Phantom) instead of timestamps or prompt previews

  Bug Fixes

  • Regenerated icon.ico with BMP-only entries to fix Windows cross-compilation in CI

  Notes

  Requires a Google Gemini API key. You will be prompted to enter one on first launch.

  Downloads

  • tatsumi_1.1.0_x64-setup.exe
    20 MiB

    2026-04-13 21:07:04 -07:00
  • Source Code (ZIP)
  • Source Code (TAR.GZ)
• v1.0.0 e87103908f

  Compare

  v1.0.0 Stable

  naomi released this 2026-04-09 20:23:43 -07:00 | 8 commits to main since this release

  Overview

  Initial release of Tatsumi — a desktop app for generating AI character art of Naomi using Google Gemini's image model. Naomi's reference sheet and character description are pre-loaded on every thread so you just describe the scene.

  Features

  Generation Modes

  • Avatar — Generate a square (1:1) portrait from a text prompt
  • Art — Generate a widescreen (16:9) art piece from a text prompt
  • Replace — Upload anime girl art and replace the character with Naomi

  Conversation Threads

  • Persistent threads saved locally between sessions
  • Follow-up messages to refine a result
  • Edit any user message and re-run the generation
  • Delete any message (removes it and everything after it)
  • Retry any model response

  Other

  • Model reasoning shown as a collapsible section per response
  • Cost per response displayed (input + output tokens)
  • Generated images downloadable with a timestamped filename
  • App version shown in the sidebar header

  Notes

  Requires a Google Gemini API key. You will be prompted to enter one on first launch.

  Downloads

  • tatsumi_1.0.0_x64-setup.exe
    20 MiB

    2026-04-09 21:36:13 -07:00
  • Source Code (ZIP)
  • Source Code (TAR.GZ)