Live data for your LlamaIndex agents.
Point LlamaIndex's MCP integration at one URL and your agent gets 94 keyless tools of fresh, current data your model was never trained on.
The call
No key, no SDK - here is the same data your LlamaIndex tools pull, straight from the keyless batch endpoint. Run it now:
curl -s -X POST https://dynamicfeed.ai/v1/batch \
-H "Content-Type: application/json" \
-d '{"calls":[
{"tool":"software_version","args":{"product":"python"}},
{"tool":"check_vulnerability","args":{"package":"lodash","version":"4.17.10","ecosystem":"npm"}}
]}'Sample response
# pip install llama-index-tools-mcp
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec
from llama_index.core.agent.workflow import FunctionAgent
from llama_index.llms.openai import OpenAI
mcp_client = BasicMCPClient("https://dynamicfeed.ai/mcp")
tools = await McpToolSpec(client=mcp_client).to_tool_list_async()
agent = FunctionAgent(tools=tools, llm=OpenAI(model="gpt-4o"))
resp = await agent.run("What is the latest stable Python, and is lodash 4.17.10 vulnerable?")Why live data
LlamaIndex agents are only as current as the data you feed them, and a base model's weights freeze on its training cutoff - it does not know today's package versions, today's CVEs, or this week's outages. Dynamic Feed exposes 94 tools over a keyless MCP endpoint (Streamable HTTP) at https://dynamicfeed.ai/mcp, so McpToolSpec turns each one into a LlamaIndex FunctionTool the agent can call at runtime. Spanning 20 domains and 78 sources, every result carries a provenance stamp - source, observed-at timestamp, freshness - and an Ed25519 signature. Be precise about what that signature means: it establishes integrity and attribution under the published key, so altered bytes fail verification. It does not independently prove time, truth, accuracy, or safety - your agent still reasons over the facts. The point is verify-not-trust: you can check the signature and trace the source instead of taking an answer on faith.
Use it for
- Load 94 live tools into a LlamaIndex FunctionAgent with McpToolSpec and zero keys
- Ground a RAG or workflow answer in a citeable, source-timed fact instead of a stale weight
- Let a coding agent check live package versions and CVEs before it recommends a dependency
- Add hazard and outage awareness - weather alerts, earthquakes, service status - to any agent workflow
FAQ
How do I connect Dynamic Feed to LlamaIndex?
Install llama-index-tools-mcp, create a BasicMCPClient pointed at https://dynamicfeed.ai/mcp, then call McpToolSpec(client).to_tool_list_async() and pass the result to a FunctionAgent or ReActAgent. The 94 tools load as standard FunctionTools - no other glue code.
Do I need an API key?
No. The MCP endpoint and the POST /v1/batch REST endpoint are both keyless and free for fair use. Just point your client at the URL.
What does the Ed25519 signature actually prove?
It establishes integrity and attribution under the published key, so altered response bytes fail verification. It does not independently prove time, truth, accuracy, or safety. Your agent still has to reason over the facts.