DSPy
Connect a DSPy program to Edison Watch using the MCP Python SDK and dspy.Tool.from_mcp_tool, keeping a stable session per conversation so data-leak protection holds across every turn.
DSPy doesn't ship a dedicated remote-MCP object - you stand up a raw MCP ClientSession over the official Python SDK's streamablehttp_client, then convert each tool with dspy.Tool.from_mcp_tool. Your connection URL carries your API key, so no auth header is needed.
pip install dspy mcpimport asyncio
import os
import dspy
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
dspy.configure(lm=dspy.LM("openai/gpt-5.6-luna"))
class Assistant(dspy.Signature):
"""Answer the user request using the available tools."""
request: str = dspy.InputField()
answer: str = dspy.OutputField()
async def main() -> None:
url = os.environ["EDISON_MCP_URL"]
async with streamablehttp_client(url) as (read, write, _):
async with ClientSession(read, write) as session:
await session.initialize()
tools = await session.list_tools()
dspy_tools = [dspy.Tool.from_mcp_tool(session, t) for t in tools.tools]
react = dspy.ReAct(Assistant, tools=dspy_tools)
result = await react.acall(request="List my available tools.")
print(result.answer)
asyncio.run(main())Set EDISON_MCP_URL to your connection URL, e.g. https://mcp.edison.watch/mcp/<your-api-key>/?client=dspy.
streamablehttp_client yields a 3-tuple (read, write, get_session_id) - not the 2-tuple you get from stdio_client. Discard the third value as _. Copying a stdio example and unpacking two values is a common error. MCP tools are async, so drive the program with react.acall(...).
Keep a stable session across turns to preserve data-leak protection
Send a stable x-edison-conversation-id header on every turn of the same conversation. That header is what keeps Edison's data-leak protection intact across a multi-turn run: Edison tracks lethal-trifecta risk per session, so if each turn looks like a brand-new session, that protection resets - and a later turn can leak data that the accumulated risk should have blocked.
Hosted clients (Claude Code, VS Code) send it automatically. For a custom DSPy program, set the header on streamablehttp_client, keyed to your own conversation or thread id:
# Reuse one stable conversation_id for every turn of the same conversation.
async with streamablehttp_client(
url,
headers={"x-edison-conversation-id": conversation_id},
) as (read, write, _):
...Without a stable x-edison-conversation-id, each connection is treated as a fresh session that starts with empty risk state - so risk accumulated on an earlier turn won't be there to block a later exfiltration. The ?client= label is only a dashboard tag, not a session key. Use a unique id per conversation (a UUID is ideal); ids are scoped to your API key, so don't reuse one string for two different conversations.
Optional: the encrypted-secrets header
For servers with zero-knowledge-encrypted secrets, pass headers to streamablehttp_client, alongside x-edison-conversation-id:
async with streamablehttp_client(
url,
headers={
"x-edison-conversation-id": conversation_id,
"x-edison-secret-key": os.environ["EDISON_SECRET_KEY"],
},
) as (read, write, _):
...