Release Notes: Unified Context Gathering System + OpenAI Model Selection

This release introduces a powerful Unified Context Gathering System that transforms how OpenAugi discovers, reviews, and processes your notes, plus flexible OpenAI model selection for all AI operations.

✨ New Features

🧠 Unified Context Gathering System

One flexible system, multiple outputs - Instead of separate rigid workflows, you now have a three-stage pipeline that gives you complete control:

1. Configure - Choose source (linked notes or recent activity), depth, filters
2. Review - See discovered notes in checkbox list, toggle individual notes on/off
3. Process - Choose to distill into atomic notes OR publish as a single blog post

New Commands

• OpenAugi: Process Notes - Process curated sets of linked notes with link traversal up to 3 levels deep
• OpenAugi: Process Recent Activity - Weekly reviews and activity summaries with time-based discovery
• OpenAugi: Save Context - Gather research without AI processing, create reference documents

Key Capabilities

• 🔗 Link Depth Traversal: Breadth-first search up to 3 levels deep (was limited to 1)
• 📋 Checkbox Review: Preview ALL discovered notes before processing, toggle individual notes
• 📊 Smart Filtering: Character limits (default 100k), folder exclusions, journal section filtering
• 📝 Dual Output Modes:
  • Distill: Atomic notes + summary (existing behavior enhanced)
  • Publish: Single polished blog post ready for sharing (NEW!)
• 💾 Raw Context Saving: Skip AI processing, just aggregate content

🤖 OpenAI Model Selection

Flexible model configuration for all AI operations:

• Preset Models: Choose from gpt-5, gpt-5-mini, gpt-5-nano (default: gpt-5)
• Custom Override: Specify any OpenAI model name (e.g., gpt-4o-2024-11-20)
• Settings Integration: Configure in Settings → OpenAugi → OpenAI Model
• No Hardcoded Models: All model references now come from configuration

🚀 Publishing System (NEW!)

Transform your notes into polished, shareable content:

• Single Blog Post Output: Conversational tone preserving your voice
• Markdown Formatting: Headers, emphasis, short paragraphs
• Frontmatter Metadata: Type, date, prompt used, source notes
• Custom Prompts: Use different "lenses" for focused processing
• Ready to Share: Copy-paste directly to blog, Discord, forums

Example Output:

---
type: published-post
published_date: 2025-10-13T14:30:00Z
prompt_used: default
status: draft
source_notes: [[Note 1]], [[Note 2]]
---

# Building a Second Brain That Actually Works

[Polished blog post content ready for publishing]

🔧 Enhanced Settings

Context Gathering Settings

• Default Link Depth: 1-3 levels (was fixed at 1)
• Default Max Characters: 100,000 (prevents token overflow)
• Filter Recent Sections by Default: Journal section filtering toggle

OpenAI Settings

• Model Selection: Dropdown with preset options
• Custom Model Override: Text input for any model name
• Published Folder: Where blog posts are saved (OpenAugi/Published)

🐛 Bug Fixes

• Journal Section Filtering: Fixed issue where date headers with additional text (e.g., ### 2025-10-13 - new) weren't being recognized and filtered properly

🔄 Backwards Compatibility

Zero breaking changes! All existing commands still work:

• Legacy commands marked appropriately in documentation
• Existing file structures unchanged
• Settings migration handled automatically
• Old workflows continue to function