Telegram Channel Processor

Python CLI for harvesting files from Telegram channels, extracting archives, deduplicating results, and producing channel-scoped outputs. It supports both channels with known static passwords and funnel channels where the archive password is written in the Telegram post itself.

Features

• Sequential channel download workflow using tdl
• Archive extraction with 7z
• Static passwords via password, password1, password2, and similar CSV columns
• Post-derived passwords via password_source=password_in_post
• Same-message password matching for funnel channels
• Cross-channel and per-channel deduplication with rdfind
• Text aggregation and sorted unique combo output
• Optional stealer-log processing into credentials.csv and autofills.csv
• Channel validation mode that can comment out inactive CSV rows

Requirements

Python

• Python 3.10+
• Optional: tqdm for progress bars

External tools

• tdl
• 7z
• rdfind
• sort

In normal processing mode all four are required. In --process-only mode, tdl is not required because the script only processes files already present in the downloads directory.

Usage

Process channels

python3 telegram_processor.py \
  --input channels.csv \
  --start 01-04-2026 \
  --end 22-04-2026

Useful options:

• --output-dir: output directory for final results
• --download-dir: directory for downloaded and extracted channel files
• --settings: path to settings.json
• --verbose: log tdl and extraction command details
• --process-only: skip Telegram downloads and process files already in --download-dir
• --auto-clean: remove processed channel directories without prompting

Check channels only

python3 telegram_processor.py \
  --input channels.csv \
  --check-channels

To comment out inactive rows in place:

python3 telegram_processor.py \
  --input channels.csv \
  --check-channels \
  --comment-missing

--check-channels does not require --start or --end.

Inactive channels are classified from tdl output as:

• not_found: channel username or target no longer exists
• inaccessible: private, forbidden, or otherwise unreachable from the current account
• error: an unexpected failure while checking

When --comment-missing is used, inactive rows are rewritten like this:

# SomeChannel,@somechannel,secret # inaccessible

Input CSV Format

The CSV must start with:

• name
• channel

Supported optional columns:

• password_source
• password, password1, password2, and any additional password columns after channel

Static-password example:

name,channel,password
Channel1,@channel1,password123
Channel2,@channel2,

Funnel-channel example:

name,channel,password_source
OnlyLogsCloud,@OnlyLogsCloud,password_in_post

password_in_post behavior

• The processor reads the exported Telegram message for each downloaded archive.
• It only accepts passwords found in the same message as the archive file.
• It looks for case-insensitive pass or password markers followed by a delimiter such as :, =, -, or whitespace.
• Supported examples include:
  • pass: 123
  • Password: @OnlyLogsCloud
  • Password FULL LOGS - @BurnCloudLogs
• Unlabeled promo links and reserve-channel links are ignored.
• If one message contains multiple archives and one matching password, that password is applied to each archive from that message.
• Static password columns are ignored when password_source=password_in_post.
• If no same-message password is found, or extraction still fails, that archive is skipped and processing continues.

Already-commented CSV rows are ignored by the channel checker.

Processing Flow

1. Load channel definitions from the CSV.
2. In normal mode, export channel messages with tdl chat export and download matching files.
3. Deduplicate downloaded files across channels with rdfind.
4. Extract archives with 7z.
5. For passworded archives, try either:
   • configured static passwords, then no password
   • or the same-message post password when password_in_post is enabled
6. Deduplicate extracted files inside each channel directory.
7. Process stealer-log outputs when archives are present.
8. Combine and sort unique text output into a *-combo.csv file.
9. Move final result files into the output directory.
10. Optionally clean channel working directories.

Output Files

Depending on the channel contents, the script may emit:

• {channel_name}-{month-year}-combo.csv
• {channel_name}-{month-year}-credentials.csv
• {channel_name}-{month-year}-autofills.csv

All final outputs are written to the configured output directory.

Configuration

settings.json controls download, extraction, sorting, and subprocess behavior. The tracked file in this repo is a sample configuration; update the stealer_log_processor.path for your machine if you use that integration.

Current sections:

• stealer_log_processor
• tdl
• sort
• archive
• processing
• logging
• subprocess

Notable keys:

• tdl.max_parallel_downloads
• tdl.export_channel_threads
• tdl.excluded_extensions
• archive.extract_patterns
• archive.supported_extensions
• archive.extract_timeout
• archive.max_parallel_extractions
• sort.temp_dir
• processing.max_workers

Notes

• channels.csv is intentionally gitignored and treated as local operator data.
• If tqdm is not installed, the script still works; it only falls back to plain logging.