v1.0.0 — First PyPI release

BrainLayer: Persistent memory for AI agents.

- 14 MCP tools (search, store, recall, think, context, regression, etc.)
- Hybrid search: semantic (bge-large-en-v1.5) + FTS5 + RRF fusion
- Local LLM enrichment pipeline (summary, tags, importance, intent)
- Session analysis: operation grouping, plan linking
- 268K+ indexed chunks across Claude Code conversations
- Full MkDocs documentation site

feat(dashboard): brain graph filters by project, source, intent (#212)

* chore(content): add .gitignore for output and one-off scripts

Ignore generated images (output/) and experimental Imagen/ComfyUI
generation scripts to keep the repo clean.

Co-Authored-By: Claude Opus 4.6 <[email protected]>

* feat(dashboard): add brain graph filters by project, source, and intent

Filter panel lets you narrow down the 3D brain view by project,
data source (claude_code/whatsapp/youtube), and intent type.
Non-matching nodes dim and shrink smoothly. Pipeline now extracts
dominant source per session from the chunks table.

Co-Authored-By: Claude Opus 4.6 <[email protected]>

* fix: address PR review - optional source type, link color with combined filters

- Make GraphNode.source optional for backward compat with old graph data
- Revert default source to "claude_code" (99.5% of sessions are CC)
- Fix linkColor to consider both filters and search simultaneously

Co-Authored-By: Claude Opus 4.6 <[email protected]>

---------

Co-authored-by: Claude Opus 4.6 <[email protected]>