Skip to main content

​
Install

pip install langchat
Requires Python 3.9 or higher.

​
Set environment variables

All providers read credentials from the environment automatically. Create a .env file: 
OPENAI_API_KEY=sk-...
PINECONE_API_KEY=pcsk-...
PINECONE_INDEX=my-index
SUPABASE_URL=https://xxxx.supabase.co
SUPABASE_KEY=eyJhbGc...

​
Your first chatbot

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

async def main():
    lc = LangChat(
        llm=OpenAI("gpt-4o-mini"),      # reads OPENAI_API_KEY
        vector_db=Pinecone("my-index"), # reads PINECONE_API_KEY
        db=Supabase(),                  # reads SUPABASE_URL + SUPABASE_KEY
    )

    response = await lc.chat(
        query="What can you help me with?",
        user_id="alice",
    )

    print(response)        # prints response text
    print(response.text)   # same — explicit access
    print(response.status) # "success" or "error"

asyncio.run(main())

​
Sync alternative

Don’t want async? Use the sync wrapper: 
from langchat import LangChat
from langchat.providers import OpenAI, Pinecone, Supabase

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

response = lc.chat_sync(query="Hello!", user_id="alice")
print(response)

​
What LangChat does automatically

When you call chat():
  1. Reformulates the question as a standalone query (resolves “it”, “that”, etc.)
  2. Searches your Pinecone index for relevant context
  3. Reranks results with Flashrank for better precision
  4. Calls the LLM with context + conversation history
  5. Saves the exchange to Supabase
  6. Returns a typed ChatResponse object
No configuration needed for any of this — it all works out of the box.

​
What you get back

chat() returns a ChatResponse dataclass:
FieldTypeDescription
textstrThe AI’s response
status"success" | "error"Whether the call succeeded
user_idstrEcho of the user ID you passed
platformstrPlatform namespace (default: "default")
response_timefloatLatency in seconds
timestampstrISO 8601 UTC timestamp
errorstr | NoneError message if status == "error"
response = await lc.chat(query="Hello", user_id="alice")

if response:                        # True when status == "success"
    print(response.text)
    print(f"Answered in {response.response_time:.2f}s")
else:
    print(f"Error: {response.error}")

​
Launch as an API server

One line to expose a full REST API with a built-in chat UI: 
from langchat.api import create_app
from langchat.providers import OpenAI, Pinecone, Supabase
import uvicorn

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

uvicorn.run(app, host="0.0.0.0", port=8000)
Open http://localhost:8000/frontend to use the built-in chat interface.

​
Next steps

Installation

Virtual environments, uv, and dependency setup

Configuration

All providers and configuration options

Document Indexing

Load PDFs, CSVs, and other documents into Pinecone

API Reference

Complete method and parameter reference