Elo — malha P2P de mensagens para agentes de IA. Zero infraestrutura.
Project description
Elo Node — P2P Message Mesh for AI Agents
Zero infrastructure. One process. One TCP port. One ed25519 key.
Elo is a decentralized P2P message mesh for communication between AI agents. No central server, no Kafka, no Redis, no NATS. Just direct TCP between nodes.
pip install elo-node
import asyncio
from elo import Node
async def main():
node = Node("my-agent", port=7878, peers=["100.91.215.113:7878"])
await node.connect()
await node.register(agents=["analyst"], tools=["web-search"])
@node.on_task
async def handle(task):
return {"result": f"processed by {node.node_id}"}
await node.run()
asyncio.run(main())
Features
- Decentralized P2P — discovery via public tracker or Kademlia DHT
- ed25519 signatures — cryptographic identity, authenticated messages
- Capabilities — publish/subscribe of agent skills across the mesh
- Relay via tracker — nodes behind NAT/Docker can communicate through a tracker
- Zero infra — no Kafka, Redis, NATS, or central server
- Native CLI —
python -m elo serve,status,init,id
CLI
python -m elo status # Node ID, hash, keys
python -m elo id # Just the node_id
python -m elo pubkey # Public key (hex + b64)
python -m elo init # Generate persistent identity
python -m elo serve # Start an interactive node
Key Concepts
peers=is required for outbound connections. Without it the node only listens.send_task()auto-fallback: direct → InterestTable → QUERY → tracker relay → NO_PEER- HELLO_ACK with known_peers: tracker shares all peers on handshake (v0.4.4+)
discover_peers_network()— QUERY broadcast across the mesh
Changelog
| Version | Highlights |
|---|---|
| 0.4.5 | Multi-response discover, tracker returns all peers on query |
| 0.4.4 | HELLO_ACK with known_peers, send_task auto-fallback tracker |
| 0.4.3 | Relay via tracker, send_task_via_tracker() |
| 0.4.0 | Initial release |
Compatibility
- Python 3.11+
- Linux, macOS, Windows
Development
git clone https://github.com/andreocc/elo
cd elo/py
pip install -e ".[dev]"
pytest
Related Projects
- Hermes Agent — autonomous agent runtime
- Honcho — persistent memory for agents
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
elo_node-0.4.8.tar.gz
(24.2 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
elo_node-0.4.8-py3-none-any.whl
(23.3 kB
view details)
File details
Details for the file elo_node-0.4.8.tar.gz.
File metadata
- Download URL: elo_node-0.4.8.tar.gz
- Upload date:
- Size: 24.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
6ddb76161c2b349e1561afac7405af8a9c76ad7cd2843ece5b1b13fc4a8f5ece
|
|
| MD5 |
ceb05a1db38c703971883ce594043116
|
|
| BLAKE2b-256 |
35d38b137bb1422350e6223e7cf8b9f3249fa884835d48f50b9121f80f0047e8
|
File details
Details for the file elo_node-0.4.8-py3-none-any.whl.
File metadata
- Download URL: elo_node-0.4.8-py3-none-any.whl
- Upload date:
- Size: 23.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.5
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
7defd564987c8c3e8586bcc36e25c65383a6e5c085d3292c0e74d7231f31395f
|
|
| MD5 |
9e9f715911d18dc6144abbc940f8f5c0
|
|
| BLAKE2b-256 |
10ddf8ac935f8d9d323cb705091ec852c5bdf79788e764aa520c503d2eea1185
|