AgentScape - An Open and Lightweight Directory for AI Agents
AgentScape is an open and lightweight, registry for exposing and discovering AI agents. It is designed for use in proof-of-concept (PoC) and MVP-stage projects.
API Base URL: https://api.agentscape.cc
Search for Agents
Query agents by name, organization, or tag.
GET /agents
curl "https://api.agentscape.cc/agents?name=contoso"
Python:
import requests
res = requests.get("https://api.agentscape.cc/agents", params={"name": "contoso"})
print(res.json())
Register an Agent Manually
POST /agents
curl -X POST https://api.agentscape.cc/agents \
-H "Content-Type: application/json" \
-d '{
"name": "My Research Agent",
"description": "Summarizes academic content using LLMs.",
"endpoint": "https://myagent.org/api",
"openapi_url": "https://myagent.org/openapi.json",
"org_website": "https://myagent.org",
"org_email": "team@myagent.org",
"tags": ["nlp", "summarization", "research"]
}'
API Reference
Method | Endpoint | Description |
---|---|---|
GET | /agents |
List or search agents |
GET | /agents/{id} |
Fetch details of a single agent |
POST | /agents |
Register an agent manually |
Purpose and Scope
AgentScape is a minimal, open registry built to support:
- Experimental deployments of AI agents during PoC/MVP phases
- Projects using LLM-based or autonomous agents with public HTTP interfaces
- Open agent interoperability in research and early-stage development
It is not designed for production or security-critical use.
Intended Audience
This project is intended for:
- AI researchers building or evaluating autonomous agents
- Developers integrating with agent orchestration frameworks (e.g. LangChain, AutoGen)
- Teams piloting multi-agent systems and looking for simple discovery tools
Open Source Code
AgentScape is open source and available on GitHub: https://github.com/ScaleUpLabs/agentscape
Contact
For collaboration or support, contact: marco.cello@scaleuplabs.vc