feat(images,videos): unified OpenAI-style image+video endpoints

fix(images): re-attach @app .post decorator to generate_images

feat(admin): unify dashboard + login design with landing/docs theme

feat(docs): branded /docs page replacing default Swagger UI

feat(models): integrate Gemini Hub (Veo 3 video + Gemini 3 image/text)

feat(providers): use dedicated gist for models_config.json

feat(providers): persist dynamic providers in GitHub Gist (models_config.json)

feat(providers): prefix-based routing with per-model custom aliases

feat(dashboard): update admin dashboard UI

fix(index): remove accidental JSON envelope appended to FastAPI init

docs(fastapi): enhance /docs page with comprehensive OpenAPI description

fix(admin): show real provider names instead of 'built-in' in models list

feat: OpenRouter-style provider prefix routing

fix: correct indentation in _save_providers no-op comment

refactor: remove Turnstile + disable local file writes (HF ephemeral runtime)

Upload folder using huggingface_hub

feat(admin): add IP-bound API key manager + gpt-image-2 endpoint

Block IP 95.5.119.10 - deny API key generation and all API access

Enable FastAPI /docs and /redoc endpoints

refactor: extract HTML templates to separate files

Add model list, security section, finalize B&W UI

Clean up remaining gradients, finalize B&W theme

Fix streaming: separate reasoning/tool_calls/content chunks, cleanup dead code, improve landing meta

Add usage to final streaming chunk for stream_options compatibility

Remove initial role chunk that confuses streaming parsers

Add debug logging for chat requests

Fix pi.dev: never forward upstream finish_reason, send reasoning as content, normalize non-streaming

Clean snapshot