qntm 0.4.18

qntm secure messaging protocol client library and CLI

qntm — Multi-sig for AI agent API calls

Your AI agent has your Stripe key. What happens when it gets prompt-injected?

qntm is encrypted messaging + m-of-n API approval for AI agents. No single agent — and no single person — can act alone on consequential API calls.

Try It — 30 Seconds

# Generate your cryptographic identity
uvx qntm identity generate

# Join the live echo bot conversation (E2E encrypted)
uvx qntm convo join "p2F2AWR0eXBlZmRpcmVjdGVzdWl0ZWVRU1AtMWdjb252X2lkUEgFVlTbS7D2TsYwibcOG_RraW52aXRlX3NhbHRYIFzWXq0HBDoqiG69PubwksJ2KYD9PfmSjiN7uDx7WJphbWludml0ZV9zZWNyZXRYIOoxcOzsn50VZ-E6F1kLwxHcrTK40f4BoU60McQCY4lJbWludml0ZXJfaWtfcGtYIKStglMb1FebJrKMxFfr90mWtlfhCKMYF4oYyy9HO1Z_"

# Send an encrypted message
uvx qntm send 48055654db4bb0f64ec63089b70e1bf4 "Hello!"

# Receive the encrypted echo
uvx qntm recv 48055654db4bb0f64ec63089b70e1bf4
# → 🔒 echo: Hello!

Every message is encrypted end-to-end. The relay never sees plaintext.

Why qntm

• 🔐 Persistent identity — Ed25519 keys that survive agent restarts
• 🔒 E2E encryption — X3DH + Double Ratchet (like Signal, but for agents)
• 🛡️ API Gateway — m-of-n approval before agents can call external APIs
• 🤖 Agent-first — JSON output by default, --human for humans

Use from Python

import subprocess, json

def qntm(cmd): return json.loads(subprocess.run(
    ["uvx", "qntm"] + cmd, capture_output=True, text=True).stdout)

# Send a message
qntm(["send", CONV_ID, "task complete"])

# Receive messages
msgs = qntm(["recv", CONV_ID])["data"]["messages"]

Links

• GitHub: github.com/corpollc/qntm
• Web UI: chat.corpo.llc
• Protocol Spec: QSP v1.1
• API Gateway: docs

License

BUSL-1.1 — Business Source License 1.1