Internal Slack AI assistant for PD company. Responds to @mentions and emoji reactions in Slack threads using GPT-4o via LangGraph.
- @mention the bot in any channel → it replies in thread
- React with the trigger emoji (default:
pidi) to any message → it replies in thread
uv synccp .env.example .envFill in .env:
| Variable | Description |
|---|---|
SLACK_BOT_TOKEN |
xoxb-... Bot User OAuth Token |
SLACK_APP_TOKEN |
xapp-... App-Level Token (Socket Mode) |
OPENAI_API_KEY |
sk-... OpenAI API key |
OPENAI_MODEL |
LLM model (default: gpt-4o) |
TRIGGER_EMOJI |
Emoji name that triggers the bot (default: pidi) |
LOG_LEVEL |
Log level (default: INFO) |
uv run python main.pyHealth check: curl http://localhost:8000/health
docker compose up --buildThe Slack app needs the following configuration:
Bot token scopes: app_mentions:read, channels:history, groups:history, chat:write, reactions:read
Event subscriptions: app_mention, reaction_added
Socket Mode: enabled (requires an App-Level Token with connections:write scope)