v0.70.0 - Experimental AI Features (agents and mcp server)
What's Changed
Full write up https://sanskar.wtf/posts/the-future-of-robyn
Experimental AI Features
First web framework with built-in AI agents and memory.
from robyn import Robyn
from robyn.ai import agent, memory
app = Robyn(__file__)
@app.post("/chat")
async def chat(request):
mem = memory(provider="inmemory", user_id="user")
ai_agent = agent(runner="simple", memory=mem)
result = await ai_agent.run(request.json().get("query"))
return {"response": result["response"]}
MCP Support
Native Model Context Protocol integration.
# MCP Resources
@app.mcp.resource("time://current")
def current_time() -> str:
return f"Current time: {datetime.now().isoformat()}"
@app.mcp.tool(
name="echo",
description="Echo back text",
input_schema={"type": "object", "properties": {"text": {"type": "string", "description": "Text to echo"}}, "required": ["text"]},
)
def echo_tool(args):
return args.get("text", "")
# MCP Prompts
@app.mcp.prompt(
name="explain_code",
description="Generate code explanation prompt",
arguments=[
{"name": "code", "description": "Code to explain", "required": True},
{"name": "language", "description": "Programming language", "required": False},
],
)
def explain_code_prompt(args):
code = args.get("code", "")
language = args.get("language", "unknown")
return f"""Please explain this {language} code:
{language}
{code}
Include:
1. What it does
2. How it works
3. Key concepts used
"""
Note
These features are very much experimental. Checkout the examples/ folder to see the new features in action.
Install: pip install robyn==0.70.0
Full Changelog: https://github.com/sparckles/Robyn/compare/v0.69.0...v0.70.0