Smartest girl in telegram.

Try it out - @sl_chatbot yourself or check out quotes - @slushaquotes

• Human-like responses in multiple messages

  
• Characters from Chub.ai

  
• Video, voice, photo, round videos, stickers input support (and more)

  
• Long-term memory and chat summary command
• Smart replies with ability to answer multiple people in one go

  
• Extensive per-chat configuration
• Undertanding of Telegram-specific features, like reply threads, via bot messages, quotes and forwards
• Optional AI telemetry for debugging and cost/prompt review

  

• deno
• gemini api key (or other supported by ai-sdk)

1. cp scripts/env.bash.example scripts/env.bash
2. set up environment variables

scripts/run.bash

• Set WIDGET_BASE_URL to your public bot URL (example: https://example.com)
• Optional web server env vars: WEB_HOST (default 0.0.0.0), WEB_PORT (default 8080)
• Build widget assets (Deno-first):

scripts/build-widget.bash

• Use /config in chat (/config global for global config)

• slusha_http_requests_total
• slusha_http_request_duration_seconds
• slusha_telegram_updates_total
• slusha_telegram_handler_errors_total
• slusha_ai_requests_total
• slusha_ai_request_duration_seconds
• slusha_ai_failures_total
• slusha_ai_finish_reason_total
• slusha_ai_tokens_total
• slusha_rate_limit_exceeded_total
• slusha_usage_downgraded_total
• slusha_process_uptime_seconds
• slusha_process_resident_memory_bytes

Add a target in your prometheus.yml:

scrape_configs:
  - job_name: slusha
    metrics_path: /metrics
    static_configs:
      - targets:
          - localhost:8080

If Slusha runs in Docker or another host, replace localhost:8080 with the reachable address.

Import the dashboard JSON from:

• grafana/slusha-observability-dashboard.json

In Grafana:

1. Go to Dashboards -> Import
2. Upload grafana/slusha-observability-dashboard.json
3. Select your Prometheus datasource when prompted

# Build and run
docker build -t slusha-bot .
docker run -d \
  --name slusha-bot \
  --restart unless-stopped \
  -e BOT_TOKEN=your_bot_token \
  -e AI_TOKEN=your_ai_token \
  slusha-bot

# Using Docker Compose
docker-compose --profile production up -d slusha-prod