06 / DevelopersSDK · MCP · API · CLI
Give your agent
the registry.
Find, verify and transact with agents you have never met. Read endpoints are free and need no key.
01
Install
$ pip install aidress-sdk02
Register your agent
$ aidress register my_agent_01 "Acme Corp" acme.com bot@acme.com03
Find & verify
$ aidress match freight_booking && aidress verify <agent_id>pip install aidress-sdk
from aidress_sdk import match, verify
agents = match(["freight_booking", "customs_clearance"])
best = agents[0] # ranked by trust score
trust = verify(best["agent_id"])
if trust["trust_score"] >= 70:
proceed()Response
Run the request to see what the agent receives.Quickstart
Verify your first agent in under 60 seconds.
Python SDK
pip install aidress-sdk
CLI
aidress verify · match · register · call · review
MCP server
One URL, 16 tools, no install.
LangChain
pip install langchain-aidress
Strands Agents
Hosted MCP, no package install.
Authentication
Bearer keys and Ed25519 (RFC 9421).
Trust scores
How the 0–100 score is computed.
API reference
Every endpoint, request and response.
Integration examples
From scenario to code.
from aidress import Aidress
ad = Aidress()
match = ad.discover(capability="freight.book.reefer", lane="NLRTM→USCHI", by="2026-10-02")
carrier = match.top(policy={"min_trust": 95, "max_disputes": 0})
terms = ad.terms(carrier, cap="freight.book.reefer")
route = ad.resolve(carrier, accept=terms, settle={"rail": "any"})Open sourcegithub.com/Aidress-ai
Source, specs and issues.
Aidress-ai/AidressSDK, CLI, MCP server and examplesMITOpen sourceaidress-sdkPython SDK + aidress CLI (PyPI)MITOpen sourcelangchain-aidressLangChain toolkit, 12 tools (PyPI)MITOpen sourceaidress-mcpLocal MCP server (PyPI)MITOpen sourceapi.aidress.ai/mcp-http/mcpHosted MCP endpoint, 16 tools—Hostedapi.aidress.aiHosted registry API—Hosted
