know index

Index all watched directories for semantic search

Overview

The index command processes all directories in the watch list, extracting text content, chunking it, and creating vector embeddings for semantic search.

Syntax

know index [OPTIONS]

Options

--log · boolean · default: "false"

Show detailed logs during indexing, including file processing information and statistics.

Alias: -l

--ext · list[string] · default: "[supported extensions]"

Comma-separated list of file extensions to index. By default, indexes all supported file types.

Alias: -e

Example: --ext .md,.txt or --ext md,txt

--glob · list[string]

Include only files matching glob patterns. Can be specified multiple times or as comma-separated values.

Alias: -g

Example: --glob "**/*.md" --glob "notes/**"

--since · string

Only index files modified since the specified time. Accepts relative time (e.g., 7d, 12h, 30m) or absolute dates (e.g., 2024-01-15).

Examples:

• 7d - Last 7 days
• 12h - Last 12 hours
• 30m - Last 30 minutes
• 2w - Last 2 weeks
• 2024-01-15 - Since January 15, 2024
• 2024-01-15T10:30:00 - Specific timestamp

--recursive · boolean · default: "true"

Scan subdirectories recursively. Use --no-recursive to scan only top-level files.

Alias: -r (enable), -R (disable)

--chunk-size · integer · default: "512"

Size of each text chunk in tokens. Larger chunks provide more context but may reduce precision.

Alias: -c

--overlap · integer · default: "50"

Number of overlapping tokens between consecutive chunks. Helps maintain context across chunk boundaries.

Alias: -o

--force · boolean · default: "false"

Clear the entire index and re-index everything from scratch.

Alias: -f

--dry · boolean · default: "false"

Simulate indexing without writing to the database. Useful for previewing what will be indexed.

--report · Path

Write a JSON report of skipped files to the specified path.

Examples

Basic indexing

know index

Indexing 2 directories
OK Total: 143 new, 89 unchanged

Index with detailed logs

know index --log

Indexing 1 directories
Extensions: ['.md', '.txt', '.py', ...]
Chunk size: 512, overlap: 50
──────────────────────── /home/user/docs ────────────────────────
✓ Processing notes.md (3 chunks)
✓ Processing readme.txt (1 chunk)
...
OK Total: 45 new, 12 unchanged

Index only markdown files

know index --ext .md

Index files modified in last 7 days

know index --since 7d

Indexing 2 directories
OK Total: 23 new, 0 unchanged

Index with glob patterns

know index --glob "**/*.md" --glob "docs/**"

Force re-index everything

know index --force

Indexing 2 directories
OK Total: 245 new, 0 unchanged

Dry run to preview indexing

know index --dry --log

Dry run: Scanning 2 directories
Extensions: ['.md', '.txt', '.py', ...]
...
Dry run: Would index 34 new, 102 unchanged

Custom chunk settings

know index --chunk-size 1024 --overlap 100

Generate skip report

know index --report skip-report.json

Indexing 2 directories
OK Total: 56 new, 189 unchanged
Report written to skip-report.json

The index command only processes files that have been modified since the last index. Use --force to rebuild the entire index from scratch.

Use --dry with --log to see exactly what files will be indexed before making any changes.

Using --force clears the entire index. Make sure this is what you want before running it.

Time Format Reference

The --since flag accepts flexible time formats:

Format	Description	Example
<N>m	Minutes ago	30m
<N>h	Hours ago	12h
<N>d	Days ago	7d
<N>w	Weeks ago	2w
YYYY-MM-DD	Absolute date	2024-01-15
ISO 8601	Full timestamp	2024-01-15T10:30:00

Prerequisites

Before running know index, you must:

1. Add at least one directory: know add <directory>
2. Verify directories are added: know dirs

See Also

• know add - Add directories to watch list
• know prune - Remove orphaned chunks
• know search - Search indexed content