# summera
Intelligent content summarisation powered by LLMs with a beautiful TUI.
Summarise webpages, PDFs, and PowerPoint presentations from the command line or
an interactive terminal UI.
## Features
- **Structured Intelligence**: Returns typed summaries with key points, conclusions, entities, and action items
- **Multiple Sources**: Summarise webpages, local PDF files, and PPTX presentations
- **Hybrid Storage**: sled for persistent storage, tantivy for full-text search
- **Provider Agnostic**: Supports Gemini and OpenAI via rstructor
- **Beautiful TUI**: Split-pane interface with summary list and scrollable detail view
- **CLI Support**: Use from the command line for scripting and automation
## Installation
### Prerequisites
- **Rust** toolchain (`cargo`)
```bash
# Clone and build
git clone https://github.com/cladam/summa.git
cd summa
cargo build --release
# The binary is at target/release/summera
```
#### Installing from crates.io
The easiest way to install `summera` is to download it from [crates.io](https://crates.io/crates/summera). You can do it
using the following command:
```bash
cargo install summera
```
If you want to update `summera` to the latest version, execute the following command:
```bash
summera update
```
## Usage
### TUI Mode (default)
Simply run `summera` to launch the interactive TUI:
```bash
summera
```
**Key bindings:**
- `o` - Open a URL or local file to summarise
- `f` - Search stored summaries
- `↑/↓` or `j/k` - Navigate summary list
- `Tab` - Switch between list and detail panes
- `PageUp/PageDown` - Scroll detail view
- `Esc` - Clear search / Cancel dialogue
- `q` - Quit
### CLI Mode
#### Summarise a webpage
```bash
summera summarise <URL>
```
Example:
```bash
summera summarise https://cladam.github.io/2025/12/22/lewin-and-devops/
```
#### Summarise a local file
Summera can extract text from **PDF** and **PPTX** files and summarise them
just like a webpage:
```bash
# PDF
summera summarise ~/Documents/quarterly-report.pdf
# PowerPoint (PPTX)
summera summarise ./slides/architecture-overview.pptx
```
> **Note:** The legacy binary `.ppt` format is not supported, only `.pptx`
> (Office Open XML).
#### View raw extracted text
Useful for inspecting what the LLM will actually see:
```bash
summera summarise <URL-or-FILE> --raw
```
#### Search stored summaries
```bash
summera search <QUERY>
```
Example:
```bash
summera search "DevOps"
```
#### List all stored summaries
```bash
summera list
```
## Configuration
On the first run, `summera` will automatically create a default configuration file at the standard location for your
operating system:
* **macOS**: `~/Library/Application Support/summera/summera.toml`
* **Linux**: `~/.config/summera/summera.toml`
* **Windows**: `%APPDATA%\summera\summera.toml`
Summera looks for configuration in the following locations (in order):
1. `./summera.toml` (current directory)
2. `summera.toml` in default config directory
If no config file exists, a default one is created automatically.
### Example `summera.toml`
```toml
[agent]
provider = "gemini" # "gemini" or "openai"
model = "gemini-2.0-flash" # Model identifier
persona = "You are a senior research assistant specialising in technical synthesis."
prompt = "Can you provide a comprehensive summary of the given text? ..."
[storage]
path = "/path/to/data" # Where to store summaries
[api]
gemini_key = "AIza..."
```
### API Keys
Use the section in `summera.toml` or set your API key as an environment variable:
```bash
# For Gemini
export GEMINI_API_KEY="your-api-key"
# For OpenAI
export OPENAI_API_KEY="your-api-key"
```
## Data Storage
Summera stores data in two locations within the configured storage path:
- **sled database**: Stores full summary data with timestamps
- **tantivy index**: Full-text search index for fast querying
Default location: `~/.local/share/summera_data/`
## Architecture
```
src/
├── main.rs # CLI entry point and argument parsing
├── lib.rs # Library exports
├── agent.rs # LLM integration via rstructor
├── config.rs # Configuration loading and management
├── reader.rs # Local file text extraction (PDF, PPTX)
├── scraper.rs # Web content extraction
├── search.rs # Tantivy full-text search
├── storage.rs # Sled persistent storage
├── summary.rs # Summary data structure
└── ui.rs # Ratatui TUI implementation
```
## Dependencies
- **rstructor**: Structured LLM outputs with schema enforcement
- **ratatui**: Terminal UI framework
- **sled**: Embedded database for storage
- **tantivy**: Full-text search engine
- **reqwest**: HTTP client for web scraping
- **scraper**: HTML parsing and content extraction
- **pdf-extract**: PDF text extraction
- **zip** / **quick-xml**: PPTX (Office Open XML) parsing
- **tokio**: Async runtime
- **clap**: CLI argument parsing
## Supported Formats
| Webpage | `http://`, `https://` | ✅ Full |
| PDF | `.pdf` | ✅ Full |
| PowerPoint (OOXML) | `.pptx` | ✅ Full |
| PowerPoint (legacy) | `.ppt` | ❌ Not supported |
## License
MIT
## Contributing
Contributions are welcome! Please feel free to open an issue or submit a pull request.