v1.11.0

v1.11.0 (2026-07-04)

This release is published under the MIT License.

Bug Fixes

• ci: Vale Docs Prose passing baseline (#244) (#270) (#270, a547a63)

Chores

• Add gitleaks pre-commit + restore vendor_spa.py (closes #278) (#285) (#285, da586dd)

• copier: Update template v2.5.3 → v2.6.0 (conflicts + lock) [red until wizard PR] (#286) (#286, 8957b1a)

• copier: Update template v2.6.0 → v2.10.2 (#299) (#299, 2e07a46)

• copier: Update to v2.5.3 (#254) (#254, 1decabb)

• deps: Bump pip from 26.0.1 to 26.1 (#224) (#224, 2b8f0ca)

• template: Apply copier template v2.2.1 (#250) (#250, 9589fe9)

• template: Apply copier template v2.3.0 (#253) (#253, a727a9e)

Documentation

• Correct PAID_PROVIDERS default in README (openai, not openai,gemini) (#272) (#272, 5281e56)

• Reference-image input closeout + prompt guidance (#262) (#269) (#269, 2ae9af6)

Features

• Gemini multi-image composition (#260) (#267) (#267, 17e9ffc)

• OpenAI gpt-image image editing and composition (#258) (#265) (#265, 0939c46)

• OpenAI inpainting masks (#261) (#268) (#268, 3225161)

• Reference-image input — foundation + Gemini single-image i2i (#257) (#263) (#263, dac4dbd)

• SD WebUI img2img + denoising strength (#259) (#266) (#266, a5bcb88)

• providers: Refresh image-model catalog for current OpenAI + Gemini models (#294) (#294, f9761b3)

Refactoring

• Rename domain modules to template scaffold layout (#282) (#282, 6fcb15a)

• cli: De-fork CLI to stock typer; drop non-canonical HTTP logging middleware (#281) (#281, f73fb57)

• config: Adopt ProjectConfig.from_env() classmethod (replaces load_config) (#280) (#280, 4f19311)

• server: Restore template server_lifespan; remove orphaned build_event_store (closes #276) (#283) (#283, 363c73c)

Testing

• Add template client fixture + smoke coverage (closes #277) (#284) (#284, 08b7c84)

---

Detailed Changes: v1.10.1...v1.11.0