Agora: language?to?protocol gateway.

https://agoraprotocol.org/docs/getting-started

There are two ways to use Agora: as a sender agent (i.e. a client) or as a receiver agent (i.e. a server). An agent can also act as both a sender and a receiver.

In this quick tutorial, we'll establish a structured communication channel between two agents:

• A LangChain agent that wants to retrieve temperature data
• A Camel agent that has access to weather data

https://github.com/agora-protocol/paper-demo

Agora is a simple cross-platform protocol that allows heterogeneous LLMs to communicate efficienly with each other. This is achieved through the power of negotiation.

In particular, Agora agents operate as follows:

1. For rare communications, they use LLMs to talk with each other in natural language
2. For frequent communications, they use LLMs to negotiate a protocol for communication, usually involving structured data (e.g. JSON)
3. Once a protocol is finalized, they use LLMs to implement routines, simple scripts (e.g. in Python) that send or receive data
4. Future communications are handled using the routines, which means that LLMs aren't required anymore

Since natural language is supported, very different agents that have never interacted before can communicate with each other, but once a common ground is established they just use routines, which are way more efficient. This enables agents to achieve at the same time efficiency, versatility and portability.

This demo showcases a network of 100 agents interacting with each other. The agents have different LLMs (OpenAI GPT-4o, Llama 3 405b, Gemini 1.5 Pro) and different DB technologies (MongoDB, SQL), but they still complete complex, multi-agent tasks with way lower costs. In a picture:
該演示展示了一個由 100 個代理組成的相互交互的網絡。這些代理具有不同的 LLM（OpenAI GPT-4o、Llama 3 405b、Gemini 1.5 Pro）和不同的數據庫技術（MongoDB、SQL），但它們仍然以更低的成本完成復雜的多代理任務。在圖片中：

You might also be interested in our HuggingFace demo.

https://blog.gopenai.com/agent-protocols-mcp-a2a-anp-and-agora-27caa1779d16

Agora Protocol (Agora)

Agora is an ongoing effort that fixes a lingering gap. Agents are good at chatting in plain language, but most execution protocols still expect rigid, fixed?field messages. Agora lets agents move fluidly between those worlds. Each provider publishes a signed Capability Card that describes a skill in clear prose, maps it to a machine?readable input–output schema and includes an endpoint for execution.

Client agents query an Agora registry with natural language — “summarise a medical paper” or “translate from Urdu to English”, and receive cards ranked by semantic match. Once a card is chosen, Agora supplies a tiny invocation skeleton in a fixed protocol such as A2A JSON RPC, letting the agents drop down to a structured channel for speed and reliability.

MCP: centralized single?agent execution. One travel agent queries flight, hotel, and weather servers directly through its local clients, so all credentials stay on?prem and the complete itinerary is stitched together in one place.

A2A: cross?department collaboration. Inside a travel company, specialised flight, hotel, weather, and activity agents accept formal Tasks from the lead agent and stream back Artifacts, letting each department own its logic while still producing a unified plan.

ANP: cross?entity, Internet?scale cooperation. Independent airline, hotel, and weather agents discover each other online, prove identity with decentralised IDs, and exchange encrypted messages, giving the travel agent trustworthy data without a central broker.

Agora: language?to?protocol gateway. Agora parses the user’s plain instruction, generates precise flight, hotel, weather, and budget protocols, and routes them to the right agents, seamlessly translating human intent into structured execution.