Loading...
LlamaIndex excels at data retrieval and RAG. Jarvis SDK adds the execution layer — your LlamaIndex agent can not only find information but also process, transform, and act on it with 700+ tools.
pip install llama-index requests
Sign up at jarvissdk.com and create an API key.
Wrap Jarvis SDK endpoints as LlamaIndex FunctionTool objects.
Include tools when creating your LlamaIndex ReActAgent or OpenAIAgent.
Copy this into your project to get started immediately.
from llama_index.core.tools import FunctionTool
from llama_index.core.agent import ReActAgent
from llama_index.llms.openai import OpenAI
import requests, json
JARVIS_API_KEY = "jsk_your_api_key_here"
def jarvis_execute(module: str, action: str, params_json: str = "{}") -> str:
"""Execute a module from Jarvis SDK's marketplace.
Args:
module: Module name (e.g., 'text-toolkit', 'hash-toolkit')
action: Action name (e.g., 'summarize', 'sha256')
params_json: JSON string of parameters
"""
resp = requests.post(
f"https://jarvissdk.com/api/v1/modules/{module}/execute",
headers={"x-api-key": JARVIS_API_KEY, "Content-Type": "application/json"},
json={"action": action, "params": json.loads(params_json)},
)
return json.dumps(resp.json().get("result", resp.json()))
def jarvis_search(query: str) -> str:
"""Search for tools in the Jarvis SDK marketplace."""
resp = requests.get(
f"https://jarvissdk.com/api/v1/catalog/search?q={query}",
headers={"x-api-key": JARVIS_API_KEY},
)
modules = resp.json().get("modules", [])[:5]
return json.dumps([{"name": m["name"], "description": m.get("description", "")} for m in modules])
# Create LlamaIndex tools
execute_tool = FunctionTool.from_defaults(fn=jarvis_execute)
search_tool = FunctionTool.from_defaults(fn=jarvis_search)
# Build agent with Jarvis SDK tools + your RAG query engine
llm = OpenAI(model="gpt-4o")
agent = ReActAgent.from_tools(
[execute_tool, search_tool], # Add your query engine tools too
llm=llm,
verbose=True,
)
response = agent.chat("Find a text processing tool and use it to count the words in 'AI agents are powerful'")
print(response)Combine LlamaIndex's RAG capabilities with Jarvis SDK's tool execution — retrieve data AND act on it in the same agent.
Jarvis SDK tools work as standard FunctionTool objects in LlamaIndex's ReAct agents for reasoning + acting loops.
Mix Jarvis SDK tools with your LlamaIndex query engines. One agent handles both data retrieval and tool execution.
Jarvis SDK's REST API works with LlamaIndex's async agent patterns for concurrent tool execution.
Yes. Pass both your query engine tools and Jarvis SDK tools to ReActAgent.from_tools(). The agent decides when to search your index vs. execute a tool.
Yes. FunctionTool objects work with both ReActAgent and OpenAIAgent — same integration code.
Use LlamaHub for data ingestion and Jarvis SDK for tool execution. They complement each other perfectly.
Free tier includes 1,000 executions/month. No credit card required.