Skip to content
Hoard Hoard

Setup Wizard

The hoard init wizard provides an interactive way to configure Hoard. This guide explains each step in detail.

Running the Wizard

Terminal window
hoard init

For a non-interactive setup with defaults:

Terminal window
hoard init --quick

Wizard Steps

Step 1: Choose Data Sources

Step 1/4: What do you want to index?


  [x] Local files (markdown, text, code)
  [ ] Agent inbox (drop folder)
  [ ] Obsidian vault
  [ ] Browser bookmarks
  [ ] Notion export (CSV/HTML)

Select the types of data you want to index. You can enable multiple sources.

Step 2: Select Folders

Step 2/4: Local files - pick folders to index


  Detected folders with documents:


  > ~/Documents/Notes          (42 files)
  > ~/Documents/Projects       (128 files)
  > ~/Obsidian                 (356 files)


  [ ] ~/Documents              (12,847 files) ⚠️ Large


  Or enter custom path: _____________

The wizard detects common document locations and shows file counts. Large folders are flagged with a warning.

Step 3: File Types

Step 3/4: File types to include


  1. Markdown
  2. Text
  3. CSV
  4. JSON
  5. YAML
  6. reStructuredText


Select file types (comma-separated) [1,2,3,4,5,6]:

Select which file extensions to index. Supported types are .md, .txt, .csv, .json, .yaml, and .rst.

Step 4/4: Enable semantic search?


  Semantic search uses AI embeddings to find conceptually
  similar content, not just keyword matches.


    Yes, enable semantic search (downloads 90MB model)
  > No, use keyword search only (faster, no download)

Vector search is optional. Without it, Hoard uses BM25 keyword search which is fast and effective for most use cases.

Background Sync Schedule

Enable background sync schedule? [Y/n]:
Sync interval (minutes) [15]:

File Watcher

Enable file watcher for live updates? [y/N]:

AI Tools (Optional)

After the main wizard completes:

Configure AI tools now? [Y/n]:

The default is Yes. If confirmed, the wizard runs hoard setup --all to configure all detected AI tools.

Orchestrator (Beta) (Optional)

If you want multi-agent workflows, run:

Terminal window
hoard orchestrate init

This creates a registration token and configures artifact storage for Orchestrator.

Post-Wizard Output

After completion:

✓ Config saved to ~/.hoard/config.yaml
✓ Database ready at ~/.hoard/hoard.db


Running initial sync...
Syncing local_files...
  Entities: 156, Chunks: 423, Tombstoned: 0, Errors: 0

If AI tool setup was selected, the setup output follows. Restart your AI tools after setup to activate the Hoard connection.

Wizard Options

OptionDescription
--quickAccept all defaults, minimal prompts
--vectorsEnable semantic search during setup
--connector <name>Add a specific connector type

Adding Sources Later

After initial setup, add more sources without re-running the full wizard:

Terminal window
# Add a folder
hoard add ~/Documents/Work


# Add Obsidian vault
hoard add --obsidian ~/Obsidian


# Add Notion export
hoard add --notion ~/Downloads/notion-export.zip


# Set inbox folder
hoard add --inbox ~/.hoard/inbox

Reconfiguring

To reconfigure:

Terminal window
hoard init

Running hoard init again will run the full wizard and overwrite the existing configuration with your new selections.

Generated Configuration

The wizard creates ~/.hoard/config.yaml:

connectors:
  local_files:
    enabled: true
    paths:
      - ~/Documents/Notes
      - ~/Documents/Projects
    include_extensions:
      - .md
      - .txt
      - .csv
      - .json
      - .yaml
    chunk_max_tokens: 400
    chunk_overlap_tokens: 50


  inbox:
    enabled: true
    path: ~/.hoard/inbox
    include_extensions:
      - .md
      - .txt
    chunk_max_tokens: 400
    chunk_overlap_tokens: 50


  obsidian:
    enabled: true
    vault_path: ~/Obsidian


vectors:
  enabled: false
  model_name: sentence-transformers/all-MiniLM-L6-v2


sync:
  interval_minutes: 15
  watcher_enabled: false
  watcher_debounce_seconds: 2


memory:
  default_ttl_days: 30
  prune_on_sync: true


server:
  host: "127.0.0.1"
  port: 19850

Next Steps