https://github.com/orangecoding/nanosearch
A web-based full-text search tool for local directories. Easy indexing of your files, then search them instantly via a browser. Without bullshit, extremely lightweight.
https://github.com/orangecoding/nanosearch
Last synced: 18 days ago
JSON representation
A web-based full-text search tool for local directories. Easy indexing of your files, then search them instantly via a browser. Without bullshit, extremely lightweight.
- Host: GitHub
- URL: https://github.com/orangecoding/nanosearch
- Owner: orangecoding
- License: apache-2.0
- Created: 2026-04-15T10:06:51.000Z (2 months ago)
- Default Branch: master
- Last Pushed: 2026-04-17T10:52:34.000Z (2 months ago)
- Last Synced: 2026-05-18T11:08:44.652Z (about 1 month ago)
- Language: JavaScript
- Homepage: https://orange-coding.net
- Size: 1.11 MB
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# nanosearch
A web-based full-text search tool for local directories. Easy indexing of your files, then search them instantly via a browser. Without bullshit, extremely lightweight.
---
## Features
- Full-text search with SQLite FTS5 (BM25 ranking, highlighted snippets)
- Incremental re-indexing, only new or changed files are processed
- OCR for scanned PDFs and images (Tesseract.js locally, Tesseract CLI in Docker)
- Supports all sorts of files, e.g. PDF, DOCX, TXT, Markdown, JPG, PNG, TIFF
- Real-time indexing progress
- Rendering your files directly in the browser if possible
---
## Screenshots
| Search | Rendering View |
| ---------------------------- | ------------------------------------ |
|  |  |
---
# Quick Start
## "Bare Metal"
**Prerequisites:** Node 22
```bash
# 1. Install dependencies + download Tesseract language data (~25 MB, once)
yarn inst
# 2. Configure
cp .env.example .env
# 4. Start the app
yarn start
```
Open `http://localhost:3000`, click **Create Index**, then search.
## Development
**Prerequisites:** Node 22
```bash
# 1. Install dependencies + download Tesseract language data (~25 MB, once)
yarn inst
# 2. Configure
cp .env.example .env
# 3. Start the backend (Terminal 1)
cd lib/backend
node --no-deprecation --watch src/server.js
# 4. Start the frontend (Terminal 2)
yarn dev
```
Open `http://localhost:5173`, click **Create Index**, then search.
---
## Docker
```bash
# 2. Configure
cp .env.example .env
# 2. Add volume mounts to docker-compose.yml for each directory in SEARCH_DIRS
# See the comment block in docker-compose.yml
# 3. Start
docker compose up -d
```
Open `http://localhost:3000`, click **Create Index**, then search.
---
## Configuration
All options are set via environment variables (or `.env` at the project root):
| Variable | Default | Description |
| ----------------------- | -------------------- | ------------------------------------------------------------------------------------ |
| `SEARCH_DIRS` | _(required)_ | Comma-separated list of directories to index |
| `DB_PATH` | `./db/nanosearch.db` | Path to the SQLite database file |
| `OCR_BACKEND` | `js` | `js` (Tesseract.js) or `cli` (Tesseract CLI) |
| `PORT` | `3000` | Backend port |
| `LOG_LEVEL` | `info` | Pino log level |
| `RESCAN_INTERVAL_HOURS` | `0` | Automatically re-scan and re-index new/changed files every X hours. `0` disables it. |
| `EXTENSIONS` | _(see .env.example)_ | Complete list of extensions to index. Nothing is indexed unless listed here. |
---
## OCR Backends
| Backend | How | When to use |
| ------- | ---------------------------------------- | ------------------------ |
| `js` | Tesseract.js (pure Node, no system deps) | Local development |
| `cli` | System `tesseract` binary | Docker (default), faster |
Both backends recognise German and English simultaneously (`deu+eng`).