Connect: LangChain4j¶
Recipe · through our MCP server · their docs checked 2026-11-17 · MCP route tested live
LangChain4j's langchain4j-mcp module connects an AI Service to any MCP server. Point its
Streamable HTTP transport at our /mcp endpoint and the model sees our tools next to its
own @Tool methods.
Our side¶
def public_app(tools: Registry, host: str, agent_token: str, reviewer_token: str) -> Any:
"""REST at /, MCP at /mcp, one bearer check for both: for a host on the internet."""
from mcp.server.transport_security import TransportSecuritySettings
from agent_adapter.server import create_app
tokens = {
agent_token: Principal("hosted-agent"), # the token the remote framework sends
reviewer_token: Principal("reviewer@example.com", scopes=frozenset({APPROVER_SCOPE})),
}
security = TransportSecuritySettings(allowed_hosts=[host], allowed_origins=[f"https://{host}"])
return create_app(
tools,
authenticate=bearer_auth(tokens),
host="0.0.0.0", # a public bind: the public-host checks apply
allowed_hosts=[host], # REST: any other Host header gets 421
mcp_options={"transport_security": security}, # MCP: the same rule
)
hosted = registry.subset(tags=["hosted"]) # no confirm tools: hosted runtimes cannot resume them
Serve public_app(hosted, …): LangChain4j does not send our resume keys back, so a
confirm tool would park on every call. In development on one machine, local_app() is
enough.
Their side¶
Add dev.langchain4j:langchain4j-mcp (a beta module: use the version that matches your
LangChain4j BOM):
import dev.langchain4j.mcp.McpToolProvider;
import dev.langchain4j.mcp.client.DefaultMcpClient;
import dev.langchain4j.mcp.client.McpClient;
import dev.langchain4j.mcp.client.transport.McpTransport;
import dev.langchain4j.mcp.client.transport.http.StreamableHttpMcpTransport;
import dev.langchain4j.service.AiServices;
McpTransport transport = StreamableHttpMcpTransport.builder()
.url("https://tools.example.com/mcp")
.customHeaders(Map.of("Authorization", "Bearer " + System.getenv("AA_AGENT_TOKEN")))
.build();
McpClient mcp = new DefaultMcpClient.Builder()
.key("shopTools")
.transport(transport)
.build();
McpToolProvider tools = McpToolProvider.builder()
.mcpClients(mcp)
.filterToolNames("shipping_rate")
.build();
Assistant assistant = AiServices.builder(Assistant.class)
.chatModel(model)
.toolProvider(tools)
.build();
| Key | Meaning |
|---|---|
StreamableHttpMcpTransport.builder().url(...) |
our /mcp endpoint |
customHeaders(...) |
carries the bearer token on every request, POST and GET |
DefaultMcpClient.Builder().key(...) |
names the client; one per server |
McpToolProvider.builder().filterToolNames(...) |
the tools the model may see |
AiServices.builder(...).toolProvider(...) |
offers the tools to your AI Service |
Close the client (mcp.close()) when the app stops. A failed call comes back as a tool
error the model can read; our ToolError messages reach it as text.
Checklist¶
| Check | Why |
|---|---|
serve hosted, or filter to tools without confirm |
LangChain4j does not send our resume keys back |
keep the default followRedirects(false) |
a redirect could send the bearer header to another host |
| the token comes from the environment or a secret store | never in source or a config file in git |
| everything on the public-host checklist |
Sources¶
- LangChain4j MCP tutorial: https://docs.langchain4j.dev/tutorials/mcp
StreamableHttpMcpTransportsource: https://github.com/langchain4j/langchain4j/tree/main/langchain4j-mcp