tg-stories exports your own Telegram Stories through Telethon, fetches viewer lists, computes engagement analytics, and renders local JSON, Markdown, and HTML reports.

The CLI is packaged as a normal Python console script and works on macOS, Linux, and Windows. Authentication is stored in the system keyring. On macOS it can also reuse credentials created by the local telegram-telethon skill.

For normal use:

pipx install telegram-stories-analytics

For local development:

python -m venv .venv
source .venv/bin/activate
python -m pip install -e .

Create api_id and api_hash at https://my.telegram.org under API Development Tools.

Then login once:

tg-stories auth login --profile personal

Check the session:

tg-stories auth status --profile personal

On macOS, if you already use the telegram-telethon skill, this tool can reuse that Keychain session:

tg-stories auth status --profile ivanopcode --auth-source telegram-telethon

Full export:

tg-stories export --profile personal --progress --open

This writes:

exports/my-telegram-stories-full.json
exports/my-telegram-stories-analysis.json
exports/my-telegram-stories-analysis.md
exports/my-telegram-stories-analysis.html

The full JSON contains each story, Telegram story link when available, metadata, counts, and a nested viewers array with viewer identity, view date, and reaction.

If you already have a full export, recompute analytics without calling Telegram:

tg-stories analyze exports/my-telegram-stories-full.json --open

Render or open HTML from an analysis JSON:

tg-stories html exports/my-telegram-stories-analysis.json --open
tg-stories open exports/my-telegram-stories-analysis.json

The generated analysis includes:

• all-time regularity and >=80% viewers;
• recent windows, by default 30, 90, and 180 days;
• sliding-window regularity;
• fastest regular viewers by median view latency;
• story-level speed buckets: 5m, 15m, 1h, 6h, and 24h;
• top viewers by absolute reactions;
• top viewers by reaction rate;
• lapsed viewers, new recent viewers, and declining viewers;
• contact, mutual contact, non-contact, and premium segments.

Important definitions:

• view_rate = viewed_story_count / story_count
• reaction_rate_per_viewed_story = reaction_count / viewed_story_count
• view_latency_seconds = viewer view_date - story date

For CI-like or temporary usage, credentials can be provided through environment variables:

export TG_STORIES_API_ID=12345
export TG_STORIES_API_HASH=...
export TG_STORIES_SESSION=...
tg-stories export --auth-source env

Do not commit exported reports. Viewer data is personal data.