LambdaChat Slackbot

A streaming socket Slackbot that responds to user inputs using an OpenAI-compatible endpoint. The bot handles multiple users and retains context until a user says "/reset".

Features

• Streaming Responses: Responses are streamed in real-time to Slack as they are generated
• Multi-User Support: Each user has their own conversation history
• Context Retention: Conversation context is maintained until explicitly reset
• Reset Command: Users can type "/reset" to clear their conversation history
• Library Implementation: Can be used as a library in other Go applications

Requirements

• Go 1.23.4 or later
• A Slack App with the following permissions:
  • app_mentions:read
  • chat:write
  • im:history
  • im:read
  • im:write
• A Slack App Token (starts with xapp-)
• A Slack Bot Token (starts with xoxb-)
• An OpenAI-compatible API endpoint (e.g., Lambda Labs API)

Installation

Using go get

go get github.com/lambda/lambdachat-slackbot

Building from Source

Clone the repository and use the provided Makefile:

git clone https://github.com/lambda/lambdachat-slackbot.git
cd lambdachat-slackbot
make init   # Create necessary directories
make build  # Build all binaries

The Makefile provides several useful targets:

• make build - Build all binaries
• make build-slackbot - Build only the slackbot binary
• make build-cli - Build only the CLI binary
• make clean - Clean up binaries
• make test - Run tests
• make run-slackbot - Run the slackbot example
• make run-direct - Run the direct example

Usage

As a Library in a Slack Bot

package main

import (
	"log"
	"os"
	"os/signal"
	"syscall"

	"github.com/lambda/lambdachat-slackbot/internal/lib"
	"github.com/sirupsen/logrus"
)

func main() {
	// Set up logging
	logger := logrus.New()
	logger.SetLevel(logrus.InfoLevel)
	logEntry := logrus.NewEntry(logger)

	// Create the SlackBot configuration
	config := lib.Config{
		LambdaChatURL: os.Getenv("LAMBDA_CHAT_URL"),
		APIKey:        os.Getenv("LAMBDA_CHAT_API_KEY"),
		Model:         os.Getenv("LAMBDA_CHAT_MODEL"),
		Persona:       os.Getenv("LAMBDA_CHAT_PERSONA"),
		SlackAppToken: os.Getenv("SLACK_APP_TOKEN"),
		SlackBotToken: os.Getenv("SLACK_BOT_TOKEN"),
		Logger:        logEntry,
	}

	// Create the SlackBot
	bot, err := lib.NewSlackBot(config)
	if err != nil {
		log.Fatalf("Failed to create SlackBot: %v", err)
	}

	// Start the bot
	if err := bot.Run(); err != nil {
		log.Fatalf("SlackBot error: %v", err)
	}
}

As a Command-Line Tool

1. Build the slackbot command:

go build -o slackbot ./cmd/slackbot

2. Run the slackbot:

./slackbot --api-key=YOUR_API_KEY --app-token=YOUR_SLACK_APP_TOKEN --bot-token=YOUR_SLACK_BOT_TOKEN

Or with environment variables:

export LAMBDA_CHAT_API_KEY=YOUR_API_KEY
export SLACK_APP_TOKEN=YOUR_SLACK_APP_TOKEN
export SLACK_BOT_TOKEN=YOUR_SLACK_BOT_TOKEN
./slackbot

Configuration

The following configuration options are available:

Option	Environment Variable	Description	Default
--api-key	LAMBDA_CHAT_API_KEY	Lambda Chat API Key	(required)
--host	LAMBDA_CHAT_URL	Lambda Chat Host	https://api.lambdalabs.com/v1
--app-token	SLACK_APP_TOKEN	Slack App Token	(required)
--bot-token	SLACK_BOT_TOKEN	Slack Bot Token	(required)
--model	LAMBDA_CHAT_MODEL	Lambda Chat Model	deepseek-llama3.3-70b
--persona	LAMBDA_CHAT_PERSONA	Lambda Chat Persona	(Bender persona)
--log-level	LOG_LEVEL	Log level (debug, info, warn, error)	info

Direct Usage with LambdaChatter

You can also use the LambdaChatter interface directly without the Slack integration:

package main

import (
	"fmt"
	"os"

	"github.com/lambda/lambdachat-slackbot/internal/lambdachat"
	"github.com/sirupsen/logrus"
)

func main() {
	// Set up logging
	logger := logrus.New()
	logEntry := logrus.NewEntry(logger)

	// Create the LambdaChatter
	lc, err := lambdachat.New(
		logEntry,
		os.Getenv("LAMBDA_CHAT_URL"),
		os.Getenv("LAMBDA_CHAT_API_KEY"),
		lambdachat.ModelDeepSeak,
		lambdachat.PersonaBender,
	)
	if err != nil {
		fmt.Printf("Failed to create LambdaChatter: %v\n", err)
		os.Exit(1)
	}

	// Use non-streaming API
	response, err := lc.Chat("user123", "Hello, how are you?")
	if err != nil {
		fmt.Printf("Error: %v\n", err)
		return
	}
	fmt.Println(response)

	// Or use streaming API
	err = lc.ChatStream("user123", "Tell me a story", os.Stdout)
	if err != nil {
		fmt.Printf("Error: %v\n", err)
	}

	// Reset conversation context
	err = lc.Reset("user123")
	if err != nil {
		fmt.Printf("Error resetting conversation: %v\n", err)
	}
}

Interacting with the Slack Bot

• Direct Messages: Send a direct message to the bot
• Mentions: Mention the bot in a channel (@YourBot Hello!)
• Reset: Type /reset to clear your conversation history

License

MIT