Skip to main content

​
Requirements

RequirementVersion
Python3.9 or higher
PineconeAccount with an index created
SupabaseEmpty project (tables created automatically)
LLM providerOpenAI, Anthropic, Gemini, Mistral, Cohere, or local Ollama

​
Install LangChat

pip install langchat
uv is a fast Python package manager that replaces pip and venv. 
curl -LsSf https://astral.sh/uv/install.sh | sh

​
Virtual environment

uv venv
source .venv/bin/activate      # macOS / Linux
.venv\Scripts\activate         # Windows
uv pip install langchat

​
Environment variables

Create a .env file in your project root. All providers read credentials automatically — you never pass keys directly in code.
.env
# ─── LLM providers (add whichever you use) ───────────────────────
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=AIza...          # also accepted: GOOGLE_API_KEY
MISTRAL_API_KEY=...
COHERE_API_KEY=...
# Ollama needs no key — runs locally

# ─── Vector database ─────────────────────────────────────────────
PINECONE_API_KEY=pcsk-...

# ─── History database ────────────────────────────────────────────
SUPABASE_URL=https://xxxx.supabase.co
SUPABASE_KEY=eyJhbGc...         # also accepted: SUPABASE_SERVICE_ROLE_KEY

# ─── Embeddings (required for document indexing) ─────────────────
# Reuses OPENAI_API_KEY — no extra variable needed
Load .env in your script with one line: 
from langchat import LangChat
LangChat.load_env()
Or use python-dotenv: 
from dotenv import load_dotenv
load_dotenv()
Add .env to your .gitignore. Never commit API keys.

​
Pinecone setup

You need a Pinecone index before LangChat can store or retrieve documents.
  1. Sign up at pinecone.io
  2. Create a Serverless index with these settings:
    • Dimensions: 3072 for text-embedding-3-large (default) or 1536 for text-embedding-3-small
    • Metric: cosine
  3. Add PINECONE_API_KEY to your .env

​
Supabase setup

LangChat creates its tables automatically — you just need a project.
  1. Create a project at supabase.com
  2. Copy Project URL and anon key (or service_role key for server-side usage)
  3. Add SUPABASE_URL and SUPABASE_KEY to your .env
On first run, LangChat creates:
  • chat_history — stores conversation messages
  • request_metrics — stores latency and error data

​
Verify installation

import langchat
print(langchat.__version__)

​
Next steps