Hello, The first final version of blindGram Assistant 1.0 is released.

changes include:

- Added text extraction from images in twenty downloadable recognition languages. Install only the languages you need and select one or more for each image.
- Added offline conversion of voice messages and round videos to text in twenty-three downloadable languages, without Telegram Premium.
- Added immediate offline image labels and an optional full AI description that can read visible text, follow the app's language, and answer follow-up questions.
- Added "Ask AI" for any text, photo, voice message, video, music or file. Questions and answers open as separate messages on an accessible conversation screen where text can be read by word, selected and copied.
- Added summaries for long messages, captions, voice messages and round videos. OpenAI can use remote speech recognition; OpenRouter, Claude and Gemini use the installed on-device voice model before sending only the transcript, and Custom providers can use either route.
- Added one shared AI settings area for blindGram and blindGram Plus, with separate controls for each feature, provider model lists and support for OpenRouter, OpenAI, Claude, Gemini and custom OpenAI-compatible addresses.
- Added seven-day conversation and result storage. Reopening the same photo, voice message or text restores its answer instead of sending, waiting for or paying for the same request again.
- Added model downloads with spoken progress, language management, optional per-feature consent before anything leaves the device, and clear provider errors when a service cannot answer.
- Added one Assistant installation for both blindGram and blindGram Plus. Installation, updates, version information and removal are available inside the blindGram Area.

@blindGram_Assistant