⚡ Priostack Agent Context Network (ACN)
Zero-setup, model-agnostic memory orchestration & multi-agent context sharing via MCP.
Priostack ACN decouples long-term memory from LLM vendor lock-in, treating persistent state as a network resource. Powered by the Model Context Protocol (MCP) over JSON-RPC, it allows AI agents to self-register, manage isolated context spaces, and securely exchange knowledge via scoped capability grants.
🔥 Key Architectural Features
- 🤖 Zero-Setup Agent Onboarding: Agents self-register via API (
noetic.register). No dashboards, UIs, or credit cards required. - 🌐 Model-Agnostic Context: Decouple agent memory from underlying LLMs (OpenAI, Anthropic, local models).
- 🤝 Multi-Agent Context Sharing: Grant or request explicit capabilities (
read,quote,fact_use,mutate) between independent agents usingnoetic.grant. - 🧠 Fact Typology: Differentiate strictly between observations (measured facts) and declarations (rules/hypotheses) to minimize hallucinations.
- 🔌 Native MCP Integration: Hook directly into Claude Desktop, Cursor, LangChain, CrewAI, or custom swarms.
🚀 Quickstart (Python)
1. Installation
pip install priostack
2. Basic Memory Workflow
from priostack import ACNClient
# Initialize client
acn = ACNClient()
# Self-register agent & establish session
token = acn.register(display_name="primary-agent")
session_id = acn.connect(token=token)
# Create an isolated memory space
space = acn.create_space(display_name="production-rules")
# Persist structured context
acn.store(
space_id="space-1",
objects=[
{"content": "Refunds over $500 require manager approval.", "type": "declaration"},
{"content": "Customer ID #402 initiated export pipeline.", "type": "observation"}
]
)
🤝 Multi-Agent Context Sharing Example
Agents can securely share context spaces without exposing full system privileges:
from priostack import ACNClient
# --- AGENT 1: Master Agent (Context Owner) ---
master = ACNClient()
master_token = master.register(display_name="master-agent")
master.connect(token=master_token)
# Create space & store data
space = master.create_space(display_name="shared-knowledge")
master.store(space_id="space-1", objects=[
{"content": "Global API rate limit set to 100 req/min.", "type": "declaration"}
])
# --- AGENT 2: Worker Agent (Consumer) ---
worker = ACNClient()
worker_token = worker.register(display_name="worker-agent")
worker_conn = worker.connect(token=worker_token)
worker_id = worker_conn.get("data", {}).get("resolvedAccount")
# --- GRANT CAPABILITIES ---
# Master agent grants read and quote permissions to worker
master.grant_access(
space_id="space-1",
subject_principal=worker_id,
rights=["read", "quote"],
world_mutation="read"
)
# Worker reconnects to sync new permissions
worker.connect(token=worker_token)
🛠️ MCP Integration (Cursor / Claude Desktop)
Add Priostack ACN directly to your mcpServers configuration:
{
"mcpServers": {
"priostack-acn": {
"command": "npx",
"args": [
"-y",
"@modelcontextprotocol/server-fetch",
"https://priostack.com/acn/rpc"
]
}
}
}
📖 API Reference (JSON-RPC)
Endpoint: https://priostack.com/acn/rpc
| Method | Description | Key Parameters |
|---|---|---|
noetic.register |
Registers a new agent autonomously | displayName |
noetic.connect |
Establishes an active session | token, maxResponseTokens |
noetic.create_space |
Instantiates an isolated context space | sessionId, displayName, defaultRights |
noetic.store |
Persists observations/declarations | sessionId, space, objects |
noetic.grant |
Grants scoped access to another agent | sessionId, space, subjectPrincipal, rights, worldMutation |
noetic.request_access |
Requests access to a remote space | sessionId, space, requestedRights |
Tools
noetic.register
Register an autonomous ACN agent.
noetic.connect
Open an authenticated ACN session.
noetic.create_space
Create a persistent context space.
noetic.store
Store declarations and observations.
noetic.query
Retrieve relevant context.
noetic.request_access
Request access to another agent's space.
noetic.grant
Grant scoped capabilities.
noetic.revoke
Revoke a capability.
noetic.receipt
Retrieve provenance/read receipts.
noetic.metrics
Inspect context usage.
📄 License
Distributed under the MIT License. See LICENSE for more information.
Release files for priostack 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| priostack-0.1.1.tar.gz | 6.3 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| priostack-0.1.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 12.1 kB
Release files / priostack-0.1.1.tar.gz
| Download URL | priostack-0.1.1.tar.gz |
|---|---|
| Size | 6.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3135f9c29a6556cbb6ee4a72306f9eaef71a06a3edc38e06dc1bba8911017155
|
|
BLAKE2b-256 checksum How to use checksums |
c9ce546d88ab905f9c43ed6cd94b8bd48e79e87d71d3b769a81ed5360c60be77
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.3
|
Release files / priostack-0.1.1-py3-none-any.whl
| Download URL | priostack-0.1.1-py3-none-any.whl |
|---|---|
| Size | 5.8 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
25a54040c759020d3bf3cd96bdf50f1ca9ef345162f023aa98a5be3528376253
|
|
BLAKE2b-256 checksum How to use checksums |
a3289b208f91cbc725558d5c37c790b8dd23d197978f51a69ce39d51e80b289e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.3
|