feat: ship Recall desktop shell and onboarding overhaul

.gitattributes

@@ -0,0 +1,2 @@
+app/.build/**/*.pkg filter=lfs diff=lfs merge=lfs -text
+app/.build/**/*.pyz filter=lfs diff=lfs merge=lfs -text

AGENTS.md

@@ -0,0 +1,30 @@
+<claude-mem-context>
+# Memory Context
+
+# [Recall/kampala] recent context, 2026-04-29 10:42pm PDT
+
+Legend: 🎯session 🔴bugfix 🟣feature 🔄refactor ✅change 🔵discovery ⚖️decision 🚨security_alert 🔐security_note
+Format: ID TIME TYPE TITLE
+Fetch details: get_observations([IDs]) | Search: mem-search skill
+
+Stats: 13 obs (5,180t read) | 214,948t work | 98% savings
+
+### Apr 20, 2026
+7 10:24a ⚖️ Recall/Mumbai App — Full-Solution Expansion Prompt Requested
+8 " 🔵 Recall/Mumbai — Existing Architecture Bottlenecks Confirmed
+9 " ✅ Plan File Repurposed as Full Product Vision Master Prompt
+12 11:54a ⚖️ Recall — Full Product Vision Defined as Agent Build Prompt
+13 " 🟣 Phase 1A — Local AI Captioning Module Designed (captioner.py)
+14 " 🟣 Phase 1B–1C — Concurrent Ingest + FastAPI Daemon Architecture Designed
+15 " 🟣 Phase 2 — Six Source Connectors Designed with BaseConnector Interface
+16 " 🟣 Phase 1D + Phase 3 — Raycast HTTP Client Rewrite and Source-Filter UX Designed
+17 " ✅ New Dependencies Added to pyproject.toml for All Phases
+### Apr 23, 2026
+36 8:10p ⚖️ Recall App UX Overhaul — Research Phase Initiated
+S9 Recall App UX Overhaul — Research Phase Initiated (Apr 23 at 8:10 PM)
+37 8:15p 🔴 Recall Branding — Removed All "Trayce" and "vector-embedded-finder" User-Facing References
+38 8:16p 🟣 recall doctor — Health-Check Subcommand Planned and Dispatched for Implementation
+39 8:17p 🟣 recall doctor — Health-Check Subcommand Implemented in CLI
+
+Access 215k tokens of past work via get_observations([IDs]) or mem-search skill.
+</claude-mem-context>

Formula/recall.rb

@@ -0,0 +1,79 @@
+class Recall < Formula
+  include Language::Python::Virtualenv
+
+  desc "Local semantic search across your files, email, calendar, and notes"
+  homepage "https://github.com/aayu22809/Recall"
+  url "https://github.com/aayu22809/Recall/archive/refs/tags/v0.3.0.tar.gz"
+  sha256 "PLACEHOLDER_SHA256"
+  license "MIT"
+
+  depends_on "python@3.11"
+
+  # Core runtime dependencies
+  resource "chromadb" do
+    url "https://files.pythonhosted.org/packages/source/c/chromadb/chromadb-0.6.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "fastapi" do
+    url "https://files.pythonhosted.org/packages/source/f/fastapi/fastapi-0.115.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "uvicorn" do
+    url "https://files.pythonhosted.org/packages/source/u/uvicorn/uvicorn-0.30.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "httpx" do
+    url "https://files.pythonhosted.org/packages/source/h/httpx/httpx-0.27.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "python-dotenv" do
+    url "https://files.pythonhosted.org/packages/source/p/python-dotenv/python_dotenv-1.0.1.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "watchdog" do
+    url "https://files.pythonhosted.org/packages/source/w/watchdog/watchdog-4.0.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "psutil" do
+    url "https://files.pythonhosted.org/packages/source/p/psutil/psutil-5.9.8.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "rich" do
+    url "https://files.pythonhosted.org/packages/source/r/rich/rich-13.7.1.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  resource "pypdf" do
+    url "https://files.pythonhosted.org/packages/source/p/pypdf/pypdf-4.1.0.tar.gz"
+    sha256 "PLACEHOLDER"
+  end
+
+  def install
+    virtualenv_install_with_resources
+  end
+
+  def post_install
+    ohai "Recall installed!"
+    ohai "Next steps:"
+    puts "  1. Run: recall doctor          # check your setup"
+    puts "  2. Run: vef-setup              # interactive setup wizard"
+    puts "  3. Run: recall start           # start the daemon"
+    puts "  4. Run: recall index ~/Documents  # index your files"
+    puts "  5. Open Raycast and search for 'Recall'"
+    puts ""
+    puts "  For embedding (pick one):"
+    puts "    Gemini (free): add GEMINI_API_KEY=... to ~/.vef/.env"
+    puts "    Ollama (local): brew install ollama && ollama pull nomic-embed-text"
+  end
+
+  test do
+    system bin/"recall", "--help"
+  end
+end

README.md

@@ -1,5 +1,5 @@
 <h1 align="center">Recall</h1>
-<p align="center"><em>Your local second brain — one search box for every file, email, calendar event, and assignment. Nothing leaves your machine.</em></p>
+<p align="center"><em>One search box for your files, email, calendar, and notes — running entirely on your Mac.</em></p>
 
 <p align="center">
   <img alt="platform" src="https://img.shields.io/badge/platform-macOS-111?style=flat-square">
@@ -9,14 +9,22 @@
   <img alt="mcp" src="https://img.shields.io/badge/MCP-compatible-8A2BE2?style=flat-square">
 </p>
 
-![System architecture](docs/figures/fig4_architecture.png)
+Recall is a background daemon that indexes your **files, Gmail, Google Calendar, Google Drive, Canvas/Schoology, Notion, and cal.ai** into a local vector database on your Mac. Search everything from Raycast, the `recall` CLI, or any MCP client. No data leaves your machine.
 
-Recall is a persistent background daemon that indexes **your files, Gmail, Google Calendar, Google Drive, cal.ai, Canvas / Schoology, and Notion** into a single local vector database. A Raycast extension, a `recall` CLI, and an MCP server all query the same in-process index over loopback HTTP.
+```
+recall search "meeting notes from last week"
+recall search "golden gate bridge photo"
+recall search "python async generator example"
+```
 
-- **Local-first.** Embeddings, vectors, and OAuth tokens live on disk. The only outbound traffic is your chosen embedding provider (or zero, if you run [Ollama](https://ollama.ai) locally).
-- **One search box, everything.** Gmail thread, PDF in `~/Downloads`, assignment posted in Canvas, meeting on your calendar — same query, same ranking.
+- **Private.** Embeddings and tokens live on disk. The only outbound call is to your embedding provider — or zero, if you run [Ollama](https://ollama.ai) locally.
 - **Fast.** Sub-200 ms semantic search over 1,000+ documents on an M-series Mac.
-- **Boring.** The daemon starts in under 4 s, ignores duplicate files via SHA-256, and gets out of the way when you're actively using the machine.
+- **Unobtrusive.** Backs off when you're actively using the machine; ignores duplicates via SHA-256 hashing.
+- **Not screen recording.** Recall indexes discrete artifacts (files, emails, events) — not your screen or microphone.
+
+### Why local-first?
+
+Cloud search tools (Notion AI, Mem, etc.) send your content to third-party AI providers for indexing and query processing. Recall does not. Your index is a folder on your disk — delete it with `rm -rf ~/.vef/chroma` to start over. No account deletion forms, no retention windows, no training on your data.
 
 <p align="center">
   <img src="docs/figures/fig3_latency.png" width="70%" alt="Search latency benchmark">

app/.build/sidecar-aarch64-apple-darwin/recall-daemon-aarch64-apple-darwin.spec

@@ -0,0 +1,51 @@
+# -*- mode: python ; coding: utf-8 -*-
+from PyInstaller.utils.hooks import collect_data_files
+from PyInstaller.utils.hooks import collect_all
+
+datas = []
+binaries = []
+hiddenimports = ['vector_embedded_finder.daemon', 'vector_embedded_finder.connectors.gmail', 'vector_embedded_finder.connectors.gcal', 'vector_embedded_finder.connectors.gdrive', 'vector_embedded_finder.connectors.calai', 'vector_embedded_finder.connectors.canvas', 'vector_embedded_finder.connectors.schoology', 'vector_embedded_finder.connectors.notion']
+datas += collect_data_files('chromadb')
+datas += collect_data_files('google')
+datas += collect_data_files('faster_whisper')
+tmp_ret = collect_all('chromadb')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('rfc3987_syntax')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+
+
+a = Analysis(
+    ['../../../vector_embedded_finder/_sidecar_entry.py'],
+    pathex=[],
+    binaries=binaries,
+    datas=datas,
+    hiddenimports=hiddenimports,
+    hookspath=[],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=[],
+    noarchive=False,
+    optimize=0,
+)
+pyz = PYZ(a.pure)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    a.binaries,
+    a.datas,
+    [],
+    name='recall-daemon-aarch64-apple-darwin',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=True,
+    upx_exclude=[],
+    runtime_tmpdir=None,
+    console=True,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch='arm64',
+    codesign_identity=None,
+    entitlements_file=None,
+)