Skip to main content

​
Minimal chatbot

The smallest working example: 
import asyncio
from langchat import LangChat
from langchat.providers import OpenAI, Pinecone, Supabase

LangChat.load_env()

lc = LangChat(
    llm=OpenAI("gpt-4o-mini"),
    vector_db=Pinecone("my-index"),
    db=Supabase(),
)

async def main():
    response = await lc.chat(query="Hello!", user_id="alice")
    print(response)

asyncio.run(main())

​
Interactive terminal chat

import asyncio
from langchat import LangChat, ChatResponse
from langchat.providers import OpenAI, Pinecone, Supabase

LangChat.load_env()

lc = LangChat(
    llm=OpenAI("gpt-4o-mini"),
    vector_db=Pinecone("my-index"),
    db=Supabase(),
)

async def main():
    user_id = "alice"
    print("LangChat — type 'exit' to quit\n")

    while True:
        query = input("You: ").strip()
        if query.lower() in ("exit", "quit", "q"):
            break
        if not query:
            continue

        response: ChatResponse = await lc.chat(query=query, user_id=user_id)

        if response:
            print(f"Bot: {response.text}")
            print(f"     [{response.response_time:.2f}s]\n")
        else:
            print(f"Bot: Error — {response.error}\n")

asyncio.run(main())

​
Sync (no async)

Use chat_sync() when you don’t want to deal with asyncio: 
from langchat import LangChat
from langchat.providers import OpenAI, Pinecone, Supabase

LangChat.load_env()

lc = LangChat(
    llm=OpenAI("gpt-4o-mini"),
    vector_db=Pinecone("my-index"),
    db=Supabase(),
)

response = lc.chat_sync(query="What is LangChat?", user_id="alice")
print(response)

​
Multiple users

Each user_id gets separate conversation history: 
async def handle_user(lc: LangChat, user_id: str, query: str):
    response = await lc.chat(query=query, user_id=user_id)
    return response.text

async def main():
    # These are completely separate conversations
    r1 = await handle_user(lc, "alice", "What's our refund policy?")
    r2 = await handle_user(lc, "bob", "How do I reset my password?")
    print(r1)
    print(r2)

​
Multiple platforms

Use platform to separate conversations from different surfaces: 
# Web app
web_response = await lc.chat(
    query="Show me my order history",
    user_id="alice",
    platform="web",
)

# Mobile app — same user, separate conversation
mobile_response = await lc.chat(
    query="Show me my order history",
    user_id="alice",
    platform="mobile",
)

​
Checking the response

ChatResponse supports bool and string protocols: 
response = await lc.chat(query="Hello", user_id="alice")

# Check success
if response:
    print("Success!")

# Print the text
print(response)         # same as print(response.text)
print(str(response))    # explicit cast

# Access all fields
print(response.text)
print(response.status)         # "success" or "error"
print(response.user_id)
print(response.platform)
print(response.response_time)  # seconds as float
print(response.timestamp)      # ISO 8601 UTC string
print(response.error)          # None on success

​
Different LLM providers

from langchat.providers import Anthropic, Gemini, Mistral, Cohere, Ollama

# Anthropic Claude
lc = LangChat(llm=Anthropic("claude-3-5-sonnet-20241022"), vector_db=..., db=...)

# Google Gemini
lc = LangChat(llm=Gemini("gemini-1.5-flash"), vector_db=..., db=...)

# Mistral
lc = LangChat(llm=Mistral("mistral-large-latest"), vector_db=..., db=...)

# Cohere
lc = LangChat(llm=Cohere("command-r-plus"), vector_db=..., db=...)

# Ollama (local, no API key)
lc = LangChat(llm=Ollama("llama3.2"), vector_db=..., db=...)