Feat/llm vision - #7

Merged

Shinitaii merged 21 commits into

mainfrom

feat/llm-vision

Jul 11, 2026

Commits on Jul 10, 2026

fix: reject unsafe/injection-prone names in property and meter-group DTOs

Shinitaii
and
claude
committed
fix: harden chatbot against disclosure and scope-creep leaks

Shinitaii
and
claude
committed
fix: add per-user rate limit on /chatbot to bound LLM spend

Shinitaii
and
claude
committed
chore: fix indexes
Shinitaii
committed
feat(api): serialize per-provider vision content in LlmClient

Shinitaii
and
claude
committed
feat(api): replace Gemini OCR with llm-config-driven vision extraction

Shinitaii
and
claude
committed
feat(llm-config): support independent chat and vision provider configs

Shinitaii
and
claude
committed
feat(ui): split chatbot and vision provider settings into separate tabs

Shinitaii
and
claude
committed
docs: update navigation docs for vision-driven OCR and dual LLM configs

Shinitaii
and
claude
committed
feat(api): add per-user photo-settings for meter-reading photo persistence

Shinitaii
and
claude
committed
feat(mobile): add OCR suggest to CaptureReadings, respect photo-settings

Shinitaii
and
claude
committed
feat(ui): merge upload/suggest into a single draggable photo box on readings

Shinitaii
and
claude
committed
docs: document photo-settings feature and mobile OCR in root CLAUDE.md

Shinitaii
and
claude
committed

Commits on Jul 11, 2026