whatsapp-asklisted
Install: claude install-skill RobertoReale/whatsapp-ask
# WhatsApp Ask Skill
Analyze exported WhatsApp conversations locally with SQLite, IDF-weighted search, and verifiable `[#ID]` message citations — using only the Python standard library.
## Locating the CLI
The pure-Python `wa` package lives alongside this `SKILL.md` (or at the root of the `whatsapp-ask` repository).
- If running from the repository root:
`python -m wa.cli <command> ...`
- If running from another working directory (after installing the skill):
`PYTHONPATH="<skill-directory>" python -m wa.cli --db "./data/wa.db" <command> ...`
Use `--db <path>` (or the `WA_DB` environment variable) to choose the SQLite database path (defaults to `data/wa.db`).
## Commands
### 1. Import a WhatsApp export (`.txt` or `.zip`)
```bash
python -m wa.cli import path/to/chat.txt
python -m wa.cli import path/to/chat.zip --name "Family Group"
```
Re-importing a chat with the same name replaces the previous copy atomically.
### 2. List imported chats
```bash
python -m wa.cli list
```
Shows each chat's ID, name, message count, and date span (`YYYY-MM-DD`).
### 3. Read a transcript (small/medium chats or narrow date ranges)
```bash
python -m wa.cli transcript --chats 1,2 --from 2026-09-01 --to 2026-09-30
```
Prints `<documents>` XML with lines formatted as `[#ID] dd/mm/yy hh:mm Sender: text` and writes the message count and estimated token count to stderr. If `--chats` is omitted, all imported chats are included.
### 4. Search large chats with IDF-ranked excerpts
When a chat se