Skip to content

Navigation Menu

Sign in
Sign up

neurobloom.ai

The open-source trust layer for AI agent coordination.

neurobloom.ai

BigTech built compute, models, and protocols. The trust layer was assumed, never encoded. PACT is that encoding — in progress, open source, available now.

PACT is the protocol. PACT-AX is the agent coordination layer. Start here.

Open protocols for AI agent collaboration — capability discovery, trust-weighted routing, and state handoffs between heterogeneous agents. Home of PACT — the Protocol for Agent Collaboration & Transfer.


Install

pip install pact-protocol # core intent protocol
pip install pact-langchain # LangChain integration
pip install pact-ax # multi-agent collaboration server
pip install pact-ax-client # Python SDK for pact-ax

30-second quickstart

from pact_ax_client import Agent
agent = Agent("my-agent", base_url="http://localhost:8000")
agent.register_capability("contract_review", description="Reviews NDAs")
decision = agent.route("contract_review")
if decision.routed:
 result = agent.handoff(decision.best_agent, state_data={"doc": "..."})
 agent.remember("contract_review", partner_id=decision.best_agent, outcome="positive")

The PACT layer map

Repo Layer What it handles
pact Core The open-source trust layer for AI agent coordination
pact-ax Agent experience Agent-to-agent collaboration primitives (server)
pact-ax-client SDK Python SDK for pact-ax — pip install pact-ax-client
pact-hx Human experience How agents show up for a single human
pact-hh Human-Human Humans collaborating through agents
pact-gx Governance Relational authority and structural dissent
pact-sx Systemic Relationships between entities and systems
pact-bridge Spine Connective tissue across the layers
pact-demos Examples Reference implementations and walkthroughs
pact-contract-testing Conformance Contract tests for PACT-compliant agents
rlp-0 Sibling Relational Ledger Protocol — computational relationships

What pact-ax-client gives you

Primitive What it does
Agent High-level façade — route, handoff, remember, vote in one object
Capabilities Register and discover agent skills
Trust Persistent, weighted trust scores between agents
Router Route tasks to the best trusted + capable agent
Episodic Memory Record and recall past interactions
Handoff / Transfer Prepare → send → receive state packets
Dead Letter Queue Park failed deliveries for retry
Consensus Weighted-vote consensus across agents

Start here

New to PACT? Read pact first.
Building with agents? pip install pact-ax-client and follow the quickstart above.
Running the server? See pact-ax.


Utility tools

Repo What it does
market-tools Market dashboards and stock screeners for US and India (Yahoo Finance)

MIT-licensed. Issues and PRs welcome across every layer.

Popular repositories Loading

  1. pact-ax pact-ax Public

    The open-source trust layer for AI agent coordination.

    Python 19 2

  2. market-tools market-tools Public

    Free, open-source market dashboards and quality stock screeners for US and India markets

    HTML 15

  3. pact pact Public

    AI agents can’t collaborate — because they don’t speak the same language. We built PACT, an open-source protocol to fix that.

    Python 5 3

  4. pact-contract-testing pact-contract-testing Public

    Pact contract testing framework

    Python 1

  5. pact-hx pact-hx Public

    PACT Human Experience Layer

    Python 1

  6. pact-demos pact-demos Public

    Python 1

Repositories

Loading
Type
Select type
Language
Select language
Sort
Select order
Showing 10 of 18 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading...

Most used topics

Loading...

AltStyle によって変換されたページ (->オリジナル) /