The Landscape
AI agents are rapidly becoming a new category of software.
Specialized agents are being created to perform specific tasks,
combine tools, access data, and provide focused capabilities.
As this ecosystem grows, agents need a common way to identify themselves,
describe what they can do, establish ownership, and become discoverable.
Today, many AI agents exist as isolated applications.
Humans can understand them through documentation and source code,
but machines need structured identity, capability, provenance,
and discovery information.
The next generation of AI will not be built around one model doing everything.
It will be built from specialized agents with different capabilities,
tools, data sources, and purposes.
Bellfre is building the identity and discovery infrastructure that allows
AI agents to identify themselves, describe their capabilities,
establish ownership, and participate in an open ecosystem.
The future of AI requires a common language for agents.
The Bellfre Difference
Bellfre starts with a simple principle:
An agent is not a model.
A model is a reasoning engine.
An agent is an application that combines capabilities and chooses
the right tool for each task.
|
v
Agent
+-----------------------+
| Capability Selection |
+-----------------------+
| |
Local Code AI Model
Tools Reasoning
A calculation should use calculation logic. A complex explanation may use an AI model. The agent determines the correct capability for the task.
The Five Pillars of Bellfre
-
1. Capability-Based Agents
Agents combine specialized capabilities including local code, tools, APIs, data sources, and AI models when appropriate. -
2. Bellfre Agent Tags (BAT)
A structured manifest that describes agent identity, capabilities, publisher information, and ecosystem participation. -
3. Bellfre Agent Clearinghouse (BAC)
A trusted registration and verification layer where publishers establish ownership, agents receive identity, and records become discoverable. -
4. Agent Manifest Standard
A common structure allowing every agent to describe itself from development through deployment and discovery. -
5. Bellfre Unique Identifier (BUI)
A permanent identity namespace providing publishers and agents with globally unique identifiers and provenance tracking.
How Bellfre Works
↓
Agent receives a BAT manifest
↓
Temporary identity assigned
↓
Publisher registers with BAC
↓
Ownership is verified
↓
Permanent BUI issued
↓
Agent appears in discovery
Discovery and Registration
Bellfre supports both publisher registration and ecosystem discovery.
Publishers can directly register their agents through the Bellfre
Agent Clearinghouse.
Bellfre can also discover existing agents, create temporary records,
and notify owners so they can claim and verify ownership.
↓
Temporary Record Created
↓
Owner Notified
↓
Owner Claims Agent
↓
Registration Completed
Building the Agent Ecosystem
Bellfre is creating the foundation for a future simple, complex and specialized agents can identify themselves, describe their capabilities, verify ownership, and work together.
The goal is not to replace AI models.
The goal is to organize the growing ecosystem around them.Featured Article
Explore why identity, manifests, provenance, and discovery will become essential as AI agents move from isolated applications into connected ecosystems.