BACK_TO_FEEDAICRIER_2
Pocket LLM v1.5.0 adds voice, OCR, camera capture
OPEN_SOURCE ↗
REDDIT · REDDIT// 4h agoPRODUCT UPDATE

Pocket LLM v1.5.0 adds voice, OCR, camera capture

Pocket LLM v1.5.0 adds a meaningful set of mobile-first upgrades to the app’s offline Android chat experience. The release brings voice input, image ingestion with OCR and vision model support, camera capture workflows with retake/crop/review, a previous-chats side panel, downloaded-model deletion to save storage, editable instructions with presets and custom prompts, UI customization, and assistant response copying.

// ANALYSIS

Strong release. This is the kind of update that pushes an offline LLM app from novelty toward daily utility because it covers input, memory, cleanup, and personalization in one pass.

  • Voice input makes the app easier to use as a hands-free assistant.
  • OCR plus Gemma vision and FastVLM support broadens it into practical multimodal capture.
  • Camera retake, crop, and review improve prompt quality before inference.
  • Previous chats and model deletion address two real UX pain points: context switching and storage pressure.
  • Editable instructions with presets/custom prompts make behavior tuning more accessible for non-technical users.
// TAGS
androidoffline-aillmmultimodalvoice-inputocrcamera-captureopen-sourcemobile-app

DISCOVERED

4h ago

2026-04-26

PUBLISHED

5h ago

2026-04-26

RELEVANCE

8/ 10

AUTHOR

100daggers_