Speakr is a personal, self-hosted web application designed for transcribing audio recordings
Python
Release v0.5.7 with GPT-5 support and custom summary prompt reprocessing
Murtaza Nasir339d6f7a508b
diff --git a/VERSION b/VERSIONindex c70db8fe..62291ae2 100644--- a/VERSION+++ b/VERSION@@ -1 +1 @@-v0.5.6\ No newline at end of file+v0.5.7diff --git a/docs/GPT5_SUPPORT.md b/docs/GPT5_SUPPORT.mddeleted file mode 100644index c5782394..00000000--- a/docs/GPT5_SUPPORT.md+++ /dev/null@@ -1,130 +0,0 @@-# GPT-5 Support in Speakr--Speakr now supports OpenAI's GPT-5 model family with proper parameter handling according to OpenAI's Chat Completions API specifications.--## Requirements--- **OpenAI Python SDK**: Version 2.2.0 or higher (updated in `requirements.txt`)-- GPT-5 support was added to the OpenAI SDK in 2025--## Overview--GPT-5 models require different API parameters compared to previous models like GPT-4. When using GPT-5 models with the official OpenAI API, Speakr automatically detects this and adjusts the API calls accordingly.--## Supported GPT-5 Models--- `gpt-5` - Best for complex reasoning, broad world knowledge, and code-heavy tasks-- `gpt-5-mini` - Cost-optimized reasoning and chat; balances speed, cost, and capability-- `gpt-5-nano` - High-throughput tasks, especially simple instruction-following-- `gpt-5-chat-latest` - Latest GPT-5 chat model--## Key Differences from GPT-4--### Unsupported Parameters-GPT-5 models **do not support** the following parameters:-- `temperature` - Replaced by `reasoning_effort` and `verbosity`-- `top_p` - Not supported-- `logprobs` - Not supported--### New GPT-5 Parameters--#### Reasoning Effort-Controls how many reasoning tokens the model generates before producing a response.--- **minimal** - Fastest responses, minimal reasoning tokens (best for simple tasks)-- **low** - Fast responses with basic reasoning-- **medium** - Balanced reasoning and speed (default, recommended)-- **high** - Maximum reasoning for complex tasks like coding and multi-step planning--#### Verbosity-Controls how many output tokens are generated.--- **low** - Concise responses-- **medium** - Balanced detail (default)-- **high** - Thorough explanations and detailed code--#### Token Limit-- Uses `max_completion_tokens` instead of `max_tokens`--## Configuration--### Environment Variables--Add these settings to your `.env` file:--```bash-# Use OpenAI API endpoint-TEXT_MODEL_BASE_URL=https://api.openai.com/v1-TEXT_MODEL_API_KEY=your_openai_api_key-TEXT_MODEL_NAME=gpt-5-mini--# GPT-5 specific parameters (optional, defaults shown)-GPT5_REASONING_EFFORT=medium-GPT5_VERBOSITY=medium-```--### Example Configurations--#### Fast Summarization (Low Cost)-```bash-TEXT_MODEL_NAME=gpt-5-nano-GPT5_REASONING_EFFORT=minimal-GPT5_VERBOSITY=low-```--#### Standard Usage (Recommended)-```bash-TEXT_MODEL_NAME=gpt-5-mini-GPT5_REASONING_EFFORT=medium-GPT5_VERBOSITY=medium-```--#### Complex Analysis (High Quality)-```bash-TEXT_MODEL_NAME=gpt-5-GPT5_REASONING_EFFORT=high-GPT5_VERBOSITY=high-```--## Automatic Detection--Speakr automatically detects when you're using:-1. A GPT-5 model (based on model name)-2. The official OpenAI API (based on base URL containing `api.openai.com`)--When both conditions are met, it automatically:-- Removes `temperature` parameter from API calls-- Adds `reasoning_effort` parameter-- Adds `verbosity` parameter-- Uses `max_completion_tokens` instead of `max_tokens`-- Logs that GPT-5 parameters are being used--## Use Cases--### Summarization-- **Fast summaries**: `gpt-5-nano` with `minimal` effort and `low` verbosity-- **Standard summaries**: `gpt-5-mini` with `medium` effort and `medium` verbosity-- **Detailed summaries**: `gpt-5` with `medium` effort and `high` verbosity--### Chat-- **Quick Q&A**: `gpt-5-mini` with `minimal` effort and `low` verbosity-- **Standard conversation**: `gpt-5-mini` with `low` effort and `medium` verbosity-- **Complex analysis**: `gpt-5` with `high` effort and `medium` verbosity--## Troubleshooting--### Error: "Unsupported parameter 'temperature'"-This means GPT-5 detection failed. Check that:-1. `TEXT_MODEL_BASE_URL` contains `api.openai.com`-2. `TEXT_MODEL_NAME` starts with `gpt-5` or is one of: `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-chat-latest`--### Error: Invalid reasoning_effort value-Valid values are: `minimal`, `low`, `medium`, `high`--### Error: Invalid verbosity value-Valid values are: `low`, `medium`, `high`--## References--- [OpenAI GPT-5 Documentation](https://platform.openai.com/docs/guides/latest-model)-- [OpenAI Chat Completions API](https://platform.openai.com/docs/api-reference/chat)diff --git a/docs/admin-guide/index.md b/docs/admin-guide/index.mdindex a9cc192f..0f0ca4b4 100644--- a/docs/admin-guide/index.md+++ b/docs/admin-guide/index.md@@ -25,14 +25,21 @@ Welcome to the Speakr Admin Guide! As an administrator, you control the heart of<p>Configure global limits, timeouts, file sizes, and system-wide behavior that affects all users.</p><a href="system-settings" class="card-link">Configure System →</a></div>-++ <div class="guide-card">+ <div class="card-icon">🤖</div>+ <h3>Model Configuration</h3>+ <p>Configure AI models for text generation, including GPT-5 support and provider selection.</p>+ <a href="model-configuration" class="card-link">Configure Models →</a>+ </div>+<div class="guide-card"><div class="card-icon">✨</div><h3>Default Prompts</h3><p>Customize AI behavior with default summary prompts that shape how content is processed.</p><a href="prompts" class="card-link">Set Prompts →</a></div>-+<div class="guide-card"><div class="card-icon">🔍</div><h3>Vector Store</h3>diff --git a/docs/admin-guide/model-configuration.md b/docs/admin-guide/model-configuration.mdnew file mode 100644index 00000000..0ac608a9--- /dev/null+++ b/docs/admin-guide/model-configuration.md@@ -0,0 +1,314 @@+# Model Configuration++This guide covers how to configure AI models for text generation in Speakr, including support for OpenAI's GPT-5 series and other language models.++## Overview++Speakr uses AI models for several key features:++- **Summary Generation**: Creating intelligent summaries of your transcriptions+- **Title Generation**: Automatically generating descriptive titles for recordings+- **Event Extraction**: Identifying calendar-worthy events from conversations+- **Interactive Chat**: Answering questions about your recordings+- **Speaker Identification**: Detecting speaker names from conversation context++These features are powered by large language models (LLMs) configured through your `.env` file.++## Basic Configuration++The text generation model is configured using three environment variables:++```bash+TEXT_MODEL_BASE_URL=https://openrouter.ai/api/v1+TEXT_MODEL_API_KEY=your_api_key_here+TEXT_MODEL_NAME=openai/gpt-4o-mini+```++### Choosing a Provider++**OpenRouter** (recommended for most users): Provides access to multiple AI models through a single API, often at competitive prices. Supports GPT-4, Claude, and many other models. Configure using `TEXT_MODEL_BASE_URL=https://openrouter.ai/api/v1`.++**OpenAI Direct**: Use OpenAI's API directly for access to their latest models including GPT-5. Configure using `TEXT_MODEL_BASE_URL=https://api.openai.com/v1`. This option is required for GPT-5 models with their specialized parameters.++**Custom Endpoints**: Speakr works with any OpenAI-compatible API endpoint, including self-hosted solutions like LocalAI, Ollama with OpenAI compatibility, or enterprise API gateways.++## GPT-5 Support++Speakr fully supports OpenAI's GPT-5 model family, automatically detecting and adjusting API parameters when you use GPT-5 models with the official OpenAI API.++### Requirements++- **OpenAI Python SDK**: Version 2.2.0 or higher (included in `requirements.txt`)+- **OpenAI API**: Must use `TEXT_MODEL_BASE_URL=https://api.openai.com/v1`+- **Valid API Key**: An OpenAI API key with GPT-5 access++### Supported GPT-5 Models++- **gpt-5**: Best for complex reasoning, broad world knowledge, and code-heavy tasks+- **gpt-5-mini**: Cost-optimized reasoning and chat; balances speed, cost, and capability+- **gpt-5-nano**: High-throughput tasks, especially simple instruction-following+- **gpt-5-chat-latest**: Latest GPT-5 chat model++### Key Differences from GPT-4++GPT-5 models use different parameters than previous models:++**Unsupported Parameters** (will cause errors if used):+- `temperature` - Replaced by `reasoning_effort` and `verbosity`+- `top_p` - Not supported+- `logprobs` - Not supported++**New GPT-5 Parameters**:++**Reasoning Effort**: Controls how many reasoning tokens the model generates before producing a response.++- **minimal**: Fastest responses, minimal reasoning tokens (best for simple tasks)+- **low**: Fast responses with basic reasoning+- **medium**: Balanced reasoning and speed (default, recommended)+- **high**: Maximum reasoning for complex tasks like coding and multi-step planning++**Verbosity**: Controls how many output tokens are generated.++- **low**: Concise responses+- **medium**: Balanced detail (default)+- **high**: Thorough explanations and detailed code++**Token Limits**: GPT-5 uses `max_completion_tokens` instead of `max_tokens`.++### Configuring GPT-5++Add these settings to your `.env` file:++```bash+# Use OpenAI API endpoint+TEXT_MODEL_BASE_URL=https://api.openai.com/v1+TEXT_MODEL_API_KEY=your_openai_api_key+TEXT_MODEL_NAME=gpt-5-mini++# GPT-5 specific parameters (optional, defaults shown)+GPT5_REASONING_EFFORT=medium+GPT5_VERBOSITY=medium+```++### GPT-5 Configuration Examples++**Fast Summarization (Low Cost)**:+```bash+TEXT_MODEL_NAME=gpt-5-nano+GPT5_REASONING_EFFORT=minimal+GPT5_VERBOSITY=low+```++**Standard Usage (Recommended)**:+```bash+TEXT_MODEL_NAME=gpt-5-mini+GPT5_REASONING_EFFORT=medium+GPT5_VERBOSITY=medium+```++**Complex Analysis (High Quality)**:+```bash+TEXT_MODEL_NAME=gpt-5+GPT5_REASONING_EFFORT=high+GPT5_VERBOSITY=high+```++### Automatic Detection++Speakr automatically detects when you're using:++1. A GPT-5 model (based on model name)+2. The official OpenAI API (based on base URL containing `api.openai.com`)++When both conditions are met, Speakr automatically:++- Removes `temperature` parameter from API calls+- Adds `reasoning_effort` parameter+- Adds `verbosity` parameter+- Uses `max_completion_tokens` instead of `max_tokens`+- Logs that GPT-5 parameters are being used++Check your logs for confirmation:+```+Using GPT-5 model: gpt-5-mini - applying GPT-5 specific parameters+```++### Using GPT-5 Through OpenRouter++If you use GPT-5 models through OpenRouter or other proxy services, the automatic GPT-5 parameter handling will **not** activate. These services typically handle parameter translation themselves, so Speakr uses standard parameters (temperature, max_tokens, etc.).++### Use Cases++**Summarization**:+- Fast summaries: `gpt-5-nano` with `minimal` effort and `low` verbosity+- Standard summaries: `gpt-5-mini` with `medium` effort and `medium` verbosity+- Detailed summaries: `gpt-5` with `medium` effort and `high` verbosity++**Chat**:+- Quick Q&A: `gpt-5-mini` with `minimal` effort and `low` verbosity+- Standard conversation: `gpt-5-mini` with `low` effort and `medium` verbosity+- Complex analysis: `gpt-5` with `high` effort and `medium` verbosity++### Troubleshooting GPT-5++**Error: "Unsupported parameter 'temperature'"**++This means GPT-5 detection failed. Check that:++1. `TEXT_MODEL_BASE_URL` contains `api.openai.com`+2. `TEXT_MODEL_NAME` starts with `gpt-5` or is one of: `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, `gpt-5-chat-latest`++**Error: "Invalid reasoning_effort value"**++Valid values are: `minimal`, `low`, `medium`, `high`++**Error: "Invalid verbosity value"**++Valid values are: `low`, `medium`, `high`++### Migrating from GPT-4 to GPT-5++1. **Update dependencies** (required for GPT-5):+ ```bash+ pip install -r requirements.txt+ ```+ This upgrades the OpenAI SDK to version 2.2.0 or higher.++2. Update your `.env` file:+ ```bash+ TEXT_MODEL_NAME=gpt-5-mini # or gpt-5, gpt-5-nano+ ```++3. Add GPT-5 parameters (optional):+ ```bash+ GPT5_REASONING_EFFORT=medium+ GPT5_VERBOSITY=medium+ ```++4. Restart Speakr:+ ```bash+ docker compose restart+ ```++5. Check logs for confirmation:+ ```+ Using GPT-5 model: gpt-5-mini - applying GPT-5 specific parameters+ ```++### Performance Considerations++- **Cost**: `gpt-5-nano` < `gpt-5-mini` < `gpt-5`+- **Speed**: `minimal` < `low` < `medium` < `high` reasoning effort+- **Quality**: Generally increases with model size and reasoning effort+- **Token usage**: Higher verbosity = more output tokens++For most use cases, we recommend:+- **Model**: `gpt-5-mini`+- **Reasoning**: `medium`+- **Verbosity**: `medium`++This provides a good balance of cost, speed, and quality.++## Model Selection Guidelines++### For Summaries++The model you choose significantly impacts summary quality:++- **GPT-4 or better**: Produces nuanced, context-aware summaries with excellent understanding of complex topics+- **GPT-5-mini**: Excellent balance of quality and cost for most summarization needs+- **GPT-3.5/4o-mini**: Budget-friendly option, suitable for straightforward content+- **Claude models**: Strong performance on structured content and technical material++### For Chat++Chat features benefit from more capable models:++- **GPT-5**: Best for complex multi-turn conversations and detailed analysis+- **GPT-5-mini**: Recommended for most chat use cases+- **Claude**: Excellent for technical discussions and code-related queries++### Cost Optimization++To reduce costs while maintaining quality:++1. **Use smaller models for simple tasks**: `gpt-5-nano` or `gpt-4o-mini` handle straightforward summaries well+2. **Adjust GPT-5 reasoning effort**: Use `minimal` or `low` for quick tasks+3. **Set token limits**: Configure `SUMMARY_MAX_TOKENS` and `CHAT_MAX_TOKENS` in your `.env`+4. **Use OpenRouter**: Often provides better rates than direct API access++### Testing Configuration++After changing model configuration:++1. Restart the Speakr container+2. Create a test recording+3. Review the generated summary and title+4. Test the chat feature+5. Monitor logs for any errors or warnings++## Environment Variables Reference++```bash+# Required: API endpoint+TEXT_MODEL_BASE_URL=https://api.openai.com/v1++# Required: API key+TEXT_MODEL_API_KEY=your_api_key_here++# Required: Model identifier+TEXT_MODEL_NAME=gpt-5-mini++# Optional: Maximum tokens for summaries (default: 8000)+SUMMARY_MAX_TOKENS=8000++# Optional: Maximum tokens for chat responses (default: 2000)+CHAT_MAX_TOKENS=2000++# GPT-5 specific (only used with GPT-5 models and OpenAI API)+GPT5_REASONING_EFFORT=medium # minimal, low, medium, high+GPT5_VERBOSITY=medium # low, medium, high+```++## Troubleshooting++### Model Not Responding++Check logs for authentication errors:+```bash+docker compose logs -f app | grep "LLM"+```++Common issues:+- Invalid API key+- Model name not available on your plan+- Rate limits exceeded+- Insufficient credits++### Poor Summary Quality++Try these adjustments:+- Upgrade to a more capable model+- Increase `SUMMARY_MAX_TOKENS`+- Review and refine [custom prompts](prompts.md)+- For GPT-5: increase reasoning effort to `medium` or `high`++### High Costs++Reduce costs with:+- Switch to smaller models (`gpt-5-nano`, `gpt-4o-mini`)+- Lower token limits+- For GPT-5: reduce reasoning effort to `minimal` or `low`+- Use OpenRouter for better rates++## Additional Resources++- [OpenAI GPT-5 Documentation](https://platform.openai.com/docs/guides/latest-model)+- [OpenAI Chat Completions API](https://platform.openai.com/docs/api-reference/chat)+- [OpenRouter Documentation](https://openrouter.ai/docs)+- [Custom Prompts Guide](prompts.md)+- [System Settings](system-settings.md)++---++Next: [Default Prompts](prompts.md) | Back to [Admin Guide](index.md)diff --git a/docs/features.md b/docs/features.mdindex b6c324e8..504e71da 100644--- a/docs/features.md+++ b/docs/features.md@@ -20,7 +20,7 @@ Transcribe content in dozens of languages with automatic detection or manual sel### Automatic Summarization-Every recording receives an AI-generated summary that captures key points, decisions, and action items. Configure this through [custom prompts](admin-guide/prompts.md). Users can also set [personal prompts](user-guide/settings.md#custom-prompts-tab) for their recordings. Summaries adapt to your content type - technical meetings get detailed technical summaries, while casual conversations receive lighter overviews. Custom prompts let you shape summaries to match your specific needs.+Every recording receives an AI-generated summary that captures key points, decisions, and action items. Configure this through [custom prompts](admin-guide/prompts.md). Users can also set [personal prompts](user-guide/settings.md#custom-prompts-tab) for their recordings. Summaries adapt to your content type - technical meetings get detailed technical summaries, while casual conversations receive lighter overviews. Custom prompts let you shape summaries to match your specific needs. For optimal results, Speakr supports [advanced AI models](admin-guide/model-configuration.md) including OpenAI's GPT-5 series with specialized parameters for reasoning depth and output verbosity.### Event Extractiondiff --git a/docs/getting-started.md b/docs/getting-started.mdindex e4380318..094306ac 100644--- a/docs/getting-started.md+++ b/docs/getting-started.md@@ -64,6 +64,8 @@ WHISPER_MODEL=whisper-1The text model can use OpenRouter for access to various AI models, or you can point it directly to OpenAI by using the same base URL and API key as your transcription service. OpenRouter provides access to multiple models including GPT-4, Claude, and others, which can be more cost-effective for text generation tasks.+For OpenAI's latest GPT-5 models (`gpt-5`, `gpt-5-mini`, `gpt-5-nano`), you must use the OpenAI API directly with `TEXT_MODEL_BASE_URL=https://api.openai.com/v1`. See the [Model Configuration Guide](admin-guide/model-configuration.md) for detailed GPT-5 setup and optimization.+### Option B: Custom ASR Endpoint Configuration> **Prerequisites:** This option requires running an additional ASR service container (`onerahmet/openai-whisper-asr-webservice`). You can either:diff --git a/docs/index.md b/docs/index.mdindex 43aab09e..a6547b6f 100644--- a/docs/index.md+++ b/docs/index.md@@ -112,17 +112,17 @@ Speakr is a powerful self-hosted transcription platform that helps you capture,## Latest Updates-!!! info "Version 0.5.6 - Latest Release"- - **Event Extraction** - Automatically identify and export calendar events from recordings- - **Transcript Templates** - Customize download formats with flexible templating system- - **Enhanced Export Options** - Download transcripts in multiple formats for different use cases- - **Improved UI** - Better mobile layouts and loading overlays across all pages+!!! info "Version 0.5.7 - Latest Release"+ - **GPT-5 Support** - Full support for OpenAI's GPT-5 model family with automatic parameter detection+ - **Custom Summary Prompts on Reprocessing** - Experiment with different prompts when regenerating summaries+ - **Enhanced Model Configuration** - Comprehensive documentation for configuring AI models+ - **Navigation Warnings** - Prevent data loss with warnings when leaving pages with unsaved work- Previous release (v0.5.5):-- - Full internationalization with multi-language support- - Enhanced audio processing at 128kbps for better accuracy- - Advanced filtering with date presets and multi-tag selection+ Previous release (v0.5.6):++ - Event extraction for automatically identifying calendar-worthy events+ - Transcript templates for customizable download formats+ - Enhanced export options and improved mobile UI## Getting Helpdiff --git a/docs/user-guide/transcripts.md b/docs/user-guide/transcripts.mdindex a37fe4b3..8a7f0d9f 100644--- a/docs/user-guide/transcripts.md+++ b/docs/user-guide/transcripts.md@@ -116,7 +116,23 @@ The summary is fully editable using a markdown editor similar to the one in theYou can export the summary in multiple ways. The copy button copies the entire summary to your clipboard in markdown format, ready to paste into emails, documents, or other applications. The download button exports the summary as a Microsoft Word (.docx) file, preserving formatting and making it easy to share with colleagues who prefer traditional documents. The downloaded file is named using the recording title and date for easy identification.-You can regenerate the summary with different settings if needed, which is useful if you want to focus on different aspects of the content or if you've made significant edits to the transcription. When regenerating, you can temporarily adjust tags to apply different prompts without permanently changing the recording's categorization.+### Reprocessing Summaries with Custom Prompts++Speakr provides powerful flexibility when you need to regenerate a summary with different instructions. The summary reprocessing feature allows you to experiment with various prompts without permanently changing your recording's configuration.++To reprocess a summary, click the reprocess button in the summary toolbar. This opens the reprocessing modal, which offers three distinct prompt options:++**Use Default Prompt**: This option applies the standard prompt hierarchy based on your current settings—using tag prompts if present, falling back to your personal prompt, then the admin default, and finally the system fallback. This is useful when you want to update the summary using your standard configuration, perhaps after editing the transcription.++**Use Prompt from Tag**: If you have created tags with custom summarization prompts in your account settings, you can select any of them to apply that tag's prompt to this specific regeneration—even if the recording doesn't have that tag applied. This is particularly powerful for experimentation: you might have a "Technical Deep Dive" tag with prompts focused on implementation details, a "Executive Summary" tag that emphasizes high-level decisions, or a "Client Presentation" tag that formats information for external audiences. You can try each one to see which produces the most useful summary for your current needs, all without changing the recording's actual tags.++When you select a tag, a preview displays the first portion of that tag's custom prompt, helping you confirm you've chosen the right one. This preview is especially helpful when you've created multiple tags with similar names but different focuses.++**Enter Custom Prompt**: For complete control, you can enter a one-time custom prompt directly in the reprocessing modal. This freeform text field lets you provide specific instructions for this single regeneration without saving the prompt anywhere. This is ideal for edge cases where you need a unique perspective on the content—perhaps "Extract all technical terms and their definitions" or "Create a summary focusing only on budget discussions" or "List all mentioned deadlines in chronological order." Your custom prompt completely overrides all other prompts, giving the AI entirely new instructions for how to process the transcription.++The custom prompt option is temporary and won't affect future regenerations or other recordings. If you find yourself using the same custom prompt repeatedly, consider creating a new tag with that prompt in your account settings so you can easily reuse it.++After selecting your prompt source and any additional options, the system processes only the summary—your transcription remains untouched. This means you can experiment freely with different summary approaches without any risk to your original content. The reprocessing respects your current transcription, including any manual edits you've made, and generates a fresh summary based solely on the prompt you've chosen.## Event Extractiondiff --git a/release_notes_v0.5.6.md b/release_notes_v0.5.6.mddeleted file mode 100644index e2bb2f9a..00000000--- a/release_notes_v0.5.6.md+++ /dev/null@@ -1,95 +0,0 @@-# Release Notes - v0.5.6--## New Features--### Event Extraction from Recordings-- **Automatic calendar event detection**: AI identifies meetings, deadlines, and appointments during summary generation-- **ICS file export**: Export events in universal calendar format compatible with all major calendar apps-- **Smart date parsing**: Understands relative references like "next Tuesday" based on recording date-- **Flexible configuration**: Toggle event extraction on/off in Account Settings under Custom Prompts-- **Intelligent defaults**: Events without specific times default to 9 AM for calendar visibility--### Transcript Templates System-- **Customizable export formats**: Create unlimited templates for different transcript use cases-- **Template variables**: Use placeholders for speaker, text, timestamps, and segment numbers-- **Format filters**: Transform variables with uppercase and SRT subtitle formatting-- **Pre-built templates**: Start with defaults for interviews, subtitles, and screenplays-- **Download flexibility**: Choose templates on-the-fly or set a default for quick exports-- **Raw export option**: Download unformatted transcripts for external processing--### Enhanced Export Capabilities-- **Template-based downloads**: Export transcripts in your preferred format-- **Multiple format support**: From simple conversations to SRT subtitles to court transcripts-- **Improved download UI**: Clean modal interface for selecting export options-- **Filename preservation**: Downloads include recording title and appropriate extensions--## Improvements--### User Interface-- **Loading overlays**: Prevent flash of unstyled content across all pages-- **Mobile tab improvements**: Better horizontal scrolling and proper sizing on small screens-- **Button consistency**: Standardized heights for copy, download, and view toggle buttons-- **Z-index fixes**: Scroll indicators no longer overlap with dropdown menus-- **Settings layout**: Improved spacing and organization in account settings tabs--### Template Management-- **Intuitive editor**: Live preview of template formatting while editing-- **Default templates button**: Quickly generate starter templates for common use cases-- **Template organization**: Clean grid layout with edit/delete capabilities-- **Comprehensive documentation**: Full guide with examples and best practices--### Event Management-- **Conditional UI**: Events tab only appears when events are detected-- **Clean event cards**: Well-organized display of event details with export buttons-- **Localization support**: Full translation of event features in all 5 supported languages--## Technical Improvements--### Database-- **Automatic migrations**: New Event model and user preferences added seamlessly-- **Template storage**: Efficient JSON storage for template definitions-- **Event relationships**: Proper foreign key constraints for data integrity--### Backend Processing-- **Efficient event extraction**: Integrated into existing summary generation pipeline-- **Template rendering**: Fast, secure template processing with Jinja2-- **Error handling**: Graceful fallbacks for malformed templates or missing data--### Frontend Enhancements-- **Dynamic template selection**: Real-time template switching in download modal-- **Responsive event display**: Adaptive layouts for different screen sizes-- **State management**: Proper handling of event extraction toggle--## Bug Fixes--- Fixed mobile view toggle buttons not matching copy button appearance-- Fixed z-index conflicts between UI elements-- Fixed FOUC (Flash of Unstyled Content) on initial page loads-- Fixed tab content spacing in settings interface-- Fixed horizontal scrolling issues in mobile tab layouts--## Configuration--### New User Settings-- **extract_events**: Boolean flag to enable/disable event extraction-- **transcript_templates**: JSON storage for custom template definitions-- **default_template_id**: User's preferred default template--### Updated Localization-- Added translations for event extraction features in all languages-- Added translations for template management interface-- Added download-related strings for transcript export options--## Migration Notes--- Existing installations automatically receive new database columns-- Event extraction is disabled by default for existing users-- No manual database migration required-- All new features are backward compatible--## Coming Next--- Bulk event export to calendar files-- Template sharing between users-- Advanced event parsing with recurring event support-- Template marketplace for community-created formats\ No newline at end of file