hoard search

The hoard search command searches your indexed content using hybrid search.

Usage

hoard search <query> [options]

Options

Option	Description
--limit <n>	Maximum results (default: 20)
--source <name>	Filter by source connector
--offset <n>	Result offset for pagination (default: 0)
--types <list>	Comma-separated result types (entity, memory)
--no-memory	Exclude memory results
--json	Output raw JSON
--config <path>	Use alternate config file

Examples

Basic Search

hoard search "meeting notes"

Limit Results

hoard search "project" --limit 5

Filter by Source

hoard search "ideas" --source obsidian

Search Only Documents

hoard search "ideas" --types entity

Search Only Memory

hoard search "preferences" --types memory

JSON Output

hoard search "query" --json

Output Format

Default Output

The CLI displays results as Rich tables, one per entity:

                        Project Notes
┏━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Type   ┃ Source       ┃ Chunk ID             ┃ Score  ┃ Content                    ┃
┡━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ entity │ local_files  │ a1b2c3d4e5f67890:2   │ 0.8700 │ ...discussed in meeting... │
└────────┴──────────────┴──────────────────────┴────────┴────────────────────────────┘

Memory results appear with Type = memory and Source = memory.

JSON Output

{
  "results": [
    {
      "result_type": "entity",
      "entity_id": "a1b2c3d4e5f6789012345678",
      "entity_title": "Project Notes",
      "source": "local_files",
      "uri": "file:///path/to/file.md",
      "chunks": [
        {
          "chunk_id": "a1b2c3d4e5f6789012345678:2",
          "content": "...discussed in the meeting notes...",
          "score": 0.87,
          "char_offset_start": 1200,
          "char_offset_end": 1850
        }
      ]
    },
    {
      "result_type": "memory",
      "entity_id": "5421d0503fadb55a413761f3745891ac",
      "entity_title": "user_preferences",
      "source": "memory",
      "memory_key": "user_preferences",
      "chunks": [
        {
          "chunk_id": "5421d0503fadb55a413761f3745891ac",
          "content": "Prefers concise responses.",
          "score": 0.71,
          "char_offset_start": null,
          "char_offset_end": null
        }
      ]
    }
  ],
  "next_cursor": "20"
}

The next_cursor field is null when there are no more results.

Search Behavior

Hybrid Search

When vectors are enabled:

1. BM25 keyword search
2. Vector semantic search
3. Reciprocal Rank Fusion merges results

When vectors are disabled:

• BM25 only (still effective for most queries)

Result Grouping

Results are grouped by entity (document), with chunks nested:

Entity A
├── Chunk 1 (score: 0.87)
├── Chunk 3 (score: 0.72)
Entity B
├── Chunk 0 (score: 0.65)

Memory entries return a single chunk per result.

Source Filters

Available source values:

• local_files
• inbox
• obsidian
• bookmarks_chrome
• bookmarks_firefox
• notion_export
• memory

No Results?

If you get no results:

1. Check sync status: hoard connectors status
2. Verify file types: Check include_extensions in config
3. Try broader query: Use fewer, more general terms
4. Run sync: hoard sync
5. Check doctor: hoard doctor

See Also

• hoard sync — Update index
• Search Concepts — How search works
• MCP Tools — Search via AI tools