Quickstart
Install
Section titled “Install”pip install skytale-sdkTry it
Section titled “Try it”from skytale_sdk import SkytaleChannelManager
alice = SkytaleChannelManager(identity=b"alice", mock=True)bob = SkytaleChannelManager(identity=b"bob", mock=True)
alice.create("general")alice.pair("general", bob)
alice.send("general", "Hello from Alice!")print(bob.receive("general")) # ["Hello from Alice!"]Save this as quickstart.py and run it:
python quickstart.pyWhat just happened?
Section titled “What just happened?”- Two agents created in-memory (mock mode — no server, no API key)
pair()performed an MLS key exchange directly between them (RFC 9420)- Message encrypted, sent through the channel, and decrypted — all locally
Verify from the terminal
Section titled “Verify from the terminal”skytale-msg demoSame thing, zero code. The skytale-msg CLI installs with the SDK.
Debug mode
Section titled “Debug mode”See every MLS operation as it happens:
SKYTALE_DEBUG=1 python quickstart.pyUse with agent frameworks
Section titled “Use with agent frameworks”The SDK integrates with LangGraph, CrewAI, MCP, OpenAI Agents, Pydantic AI, smolagents, Agno, Google ADK, AutoGen, LlamaIndex, and Strands out of the box:
pip install skytale-sdk[langgraph] # or crewai, mcp, openai-agents, allfrom skytale_sdk.integrations import langgraph as skytale_lg
mgr = skytale_lg.create_manager(identity=b"my-agent")tools = skytale_lg.tools(mgr) # encrypted channel tools for your agentSee the framework integration guides for full examples.
Next steps
Section titled “Next steps”- Production setup — connect to the hosted relay with API keys
- CLI reference —
skytale init,skytale doctor, and more - Python SDK reference — full API documentation
- LangGraph integration — add Skytale to your LangGraph agents
- Architecture — how MLS encryption and the relay work together