# LangChain / LangGraph

> Connect Piqline's MCP server to LangChain / LangGraph (LangChain): the block to copy, where it goes, and the API key in a header.

Source: https://piqline.com/docs/api/clients/langchain

Piqline reads public sources, dates what it sees and hands it to agents with the proof. This page connects its MCP server to LangChain / LangGraph.

## LangChain / LangGraph

**Where:** Python

Piqline tools become LangChain tools.

```python
import os
from langchain_mcp_adapters.client import MultiServerMCPClient

client = MultiServerMCPClient({
    "piqline": {
        "transport": "http",
        "url": "https://piqline.com/api/mcp",
        "headers": {"Authorization": f"Bearer {os.environ['PIQLINE_API_KEY']}"},
    }
})
tools = await client.get_tools()
```

Source: [github.com](https://github.com/langchain-ai/langchain-mcp-adapters), read on 20 Nov 2026 (block from the official documentation). Tested with a real key on 2026-11-20.

## Next

Ask the agent, for example: “Who published first on this subject?”, “Which companies pay to hire in France this week?”, or “Read this page and tell me since when it exists”.

- [Six things Piqline does for an agent](/assistants)
- [All clients](/docs/api/clients) · [The MCP server](/docs/api/mcp) · [Create a key](/docs/api/authentication)
