v0.5.6

Key Highlights

• Integration with Twilio for phone calls and OpenAI Agents SDK.
• Docker/MCP fixes resolving HTTP 405 errors.
• Ability to replace reference samples for voice clones.
• Audiobook chapters survive power-off and data-folder moves.
• Malayalam number verbalization and chapter cue sheet export.

New Features

• Edit saved voice clone reference samples.
• OpenAI Agents integration with copyable snippet.
• Export chapter cue sheets (HH:MM:SS<Tab>Title).
• Copyable setup for Codex CLI, MCP clients, and Docker/GHCR.
• Transcriptions pass `language`, `prompt`, and `temperature` to Whisper engines.
• Speech `pcm` is 24 kHz; `aac` and `opus` return real formats.

Full Release Notes

## What's Changed
* Make max file size configurable via SEMBLE_MAX_FILE_BYTES and warn on skipped files by @TommyC81 in https://github.com/MinishLab/semble/pull/252
* fix: permission issues in sandboxes by @stephantul in https://github.com/MinishLab/semble/pull/256
* fix: Double embedding pass on fresh index by @Pringled in https://github.com/MinishLab/semble/pull/260

## New Contributors
* @TommyC81 made their first contribution in https://github.com/MinishLab/semble/pull/252

**Full Changelog**: https://github.com/MinishLab/semble/compare/v0.5.5...v0.5.6

Related Repositories