Mirror-Log

An append-only event log for capturing thoughts, notes, and data you do not want to lose.

mirror-log is local-first, SQLite-backed, and designed to be boring in the best way: easy to inspect, easy to script, and hard to accidentally lose context.

Version 0.1.6 includes semantic embedding support and enhanced iteration tracking capabilities.

Quick Start

# Add one event
mirror-log add "Overhead allocation needs review" --source journal

# Add from file
mirror-log add-file notes.md --source meetings

# Bulk import from stdin (one line = one event)
cat ideas.txt | mirror-log stdin --source ideas

# Show recent
mirror-log show --last 10

# Search full events
mirror-log search "overhead"

# Search chunked content
mirror-log search "allocation" --chunks

# Ingestion stats (total/unique/duplicates)
mirror-log stats

# Database summary
mirror-log info

# Integrity verification (hash + relational checks)
mirror-log verify

# Generate embeddings for events (coming in next version)
# mirror-log embed --source journal

# Search similar events using embeddings (coming in next version)
# mirror-log search-similar "overhead allocation" --limit 5

Installation

# Clone and build
git clone https://github.com/CromboJambo/mirror-log
cd mirror-log
cargo build --release

# Binary location: target/release/mirror-log

# Or install locally
cargo install --path .

Documentation

User Guide - Comprehensive documentation with examples, advanced features, and best practices
API Documentation - Library usage for programmatic access

Core Principles

Append-only: Events are never updated or deleted
SQLite is source of truth: Your data stays local and inspectable
No hidden layers: Direct SQL remains first-class
Source-aware logging: Every event tracks where it came from
Semantic Embedding Support: Events can be enriched with vector embeddings for similarity search

Data Model

Main Tables

Events Table

id TEXT PRIMARY KEY (UUID)
timestamp INTEGER NOT NULL (event timestamp)
source TEXT NOT NULL
content TEXT NOT NULL
meta TEXT NULL
ingested_at INTEGER NOT NULL
content_hash TEXT NULL (SHA256 for dedupe analytics)

Chunks Table

Stores chunked slices of event content
Used by search --chunks and large-content workflows

Direct SQLite Access

sqlite3 mirror.db

# Query events
SELECT datetime(timestamp, 'unixepoch'), source, content
FROM events
ORDER BY timestamp DESC
LIMIT 10;

# Count by source
SELECT source, COUNT(*)
FROM events
GROUP BY source
ORDER BY COUNT(*) DESC;

# Get stats
SELECT COUNT(*) AS total,
       COUNT(DISTINCT content_hash) AS unique_events
FROM events;

# View embeddings (coming in next version)
# SELECT e.id, e.content, emb.embedding
# FROM events e
# JOIN event_embeddings emb ON e.id = emb.event_id
# LIMIT 10;

Development

cargo fmt
cargo clippy --all-targets --all-features -- -D warnings
cargo test

License

AGPL-3.0-or-later. See LICENSE.

mirror-log 0.1.6