An Agent Communication Language (ACL) is a standardized formal language that defines the syntax, semantics, and pragmatics of messages exchanged between autonomous agents to enable interoperable knowledge sharing and coordination. It provides a shared vocabulary and a set of speech acts—such as inform, request, or propose—that allow heterogeneous agents, potentially built on different frameworks, to understand each other's intentions and collaborate effectively within a multi-agent system (MAS).
Glossary
Agent Communication Language (ACL)

What is Agent Communication Language (ACL)?
A formal protocol enabling structured, semantic communication between autonomous software agents.
Prominent implementations include the Foundation for Intelligent Physical Agents (FIPA) ACL and the earlier Knowledge Query and Manipulation Language (KQML). These languages separate the communicative intent (the performative) from the content language (e.g., Prolog, SQL) and the ontology defining domain concepts. This separation is critical for enabling agent interoperability and complex agent coordination patterns like negotiation and cooperative problem-solving in distributed environments.
Core Components of an ACL
An Agent Communication Language (ACL) is a formal, standardized language that defines the syntax, semantics, and pragmatics of messages exchanged between autonomous agents. Its core components ensure agents can interoperate, share knowledge, and coordinate actions unambiguously.
Message Structure (Syntax)
The syntax defines the formal grammar and structure of an ACL message. It specifies the required and optional fields that constitute a valid message, ensuring parsability. A typical ACL message, as seen in standards like FIPA ACL, includes:
- Performative: The type of communicative act (e.g.,
inform,request,propose). - Sender/Receiver: The unique identifiers of the participating agents.
- Content: The core information payload of the message.
- Language & Ontology: References specifying how to interpret the content.
- Protocol & Conversation ID: Fields for managing multi-message dialogues. This rigid structure allows any compliant agent to parse and validate incoming messages.
Communicative Acts (Semantics)
The semantics define the precise meaning of message types, known as communicative acts or performatives. This is the core of an ACL, dictating the intended effect of a message on the receiver's mental state (beliefs, desires, intentions). Key performatives include:
inform: Asserts a proposition the sender believes to be true.request: Asks the receiver to perform an action.cfp(Call for Proposals): Initiates a negotiation.propose,accept-proposal,reject-proposal: Used in contract-net protocols.query-ref: Asks for the value of a referential expression. The semantics ensure that aninformmessage is treated as a statement of fact, while arequestis treated as an attempt to influence action, enabling predictable interactions.
Content Language & Ontology
An ACL separates the communication layer from the content language and ontology. The ACL defines how to communicate, while these components define what is being communicated.
- Content Language: The formal language (e.g., KIF, FIPA-SL, RDF, JSON-LD) used to express the propositions, actions, or objects within the message's
contentfield. - Ontology: A shared, machine-readable specification of concepts, properties, and relationships in a domain (e.g., e-commerce, logistics). It provides the vocabulary for the content, ensuring all agents interpret terms like
'DeliveryDate'or'AuctionBid'identically. This separation allows an ACL to be domain-agnostic, reusable across different applications by swapping the underlying ontology.
Interaction Protocols
Interaction protocols are predefined, structured sequences of communicative acts that govern complex, multi-turn conversations between agents. They provide a shared blueprint for achieving common coordination tasks. Standard protocols include:
- Request Protocol: A simple
requestfollowed by anagree/refuseand a subsequentinform(result) orfailure. - Contract Net Protocol: A one-to-many negotiation for task allocation using
cfp,propose,accept-proposal, andreject-proposal. - Iterated Contract Net: An extension for multi-round bidding.
- Auction Protocols: Such as English or Dutch auctions. Protocols manage conversation state, timeouts, and legal message sequences, moving beyond single-message exchanges to coordinated workflows.
Transport & Envelope
The transport and message envelope components handle the physical delivery of ACL messages across a network. While the ACL defines the logical message, the envelope contains the necessary metadata for routing and delivery over specific transport mechanisms.
- Envelope: Wraps the ACL message, containing lower-level addressing, routing information, and transport-specific details (e.g., encoding, encryption flags). In FIPA ACL, the envelope is a separate but associated structure.
- Transport Mechanisms: Define how envelope-wrapped messages are sent (e.g., HTTP, IIOP, MQTT, gRPC). The FIPA Agent Message Transport Service (MTS) specification standardizes this layer. This abstraction allows ACL messages to be exchanged over various network protocols without altering their semantic meaning.
Pragmatics & Rationality
Pragmatics are the conventions and assumptions that govern the use of the language in a social context. They define the pre-conditions and expected outcomes of communicative acts, often grounded in a theory of agency like the Belief-Desire-Intention (BDI) model. Key principles include:
- Sincerity: An agent sending an
informshould believe the content to be true. - Rationality: An agent should only
requestan action it believes the receiver is capable of performing. - Cooperation: Agents are assumed to be participating in good faith to achieve the goals of the interaction. These pragmatic rules are not enforced by syntax but are essential for predictable, efficient, and trustworthy multi-agent interactions. They represent the social layer of agent communication.
Major ACL Standards: KQML vs. FIPA ACL
A technical comparison of the two primary formal languages for enabling interoperable communication between autonomous agents in a multi-agent system.
| Feature | KQML (Knowledge Query and Manipulation Language) | FIPA ACL (Foundation for Intelligent Physical Agents Agent Communication Language) |
|---|---|---|
Primary Origin & Era | DARPA/University-based research (early 1990s) | International standards body FIPA (late 1990s/2000s) |
Core Design Philosophy | Extensible performative-based messaging for knowledge sharing | Formal, speech-act based communication grounded in agent mental states |
Standardized Semantic Model | ||
Standardized Content Language Syntax | ||
Standardized Interaction Protocols | ||
Standardized Agent Management | ||
Primary Message Structure | Layered: Communication layer, Message layer, Content layer | Structured envelope with defined parameters (e.g., :sender, :receiver, :content, :ontology) |
Formal Semantics Foundation | Varied, implementation-dependent | Explicitly defined using semantic frames (feasibility preconditions, rational effect) |
Typical Use Case | Academic research, early agent systems, flexible knowledge exchange | Industrial and commercial multi-agent systems requiring strict interoperability |
Modern Framework Adoption | Limited; historical influence | Widely adopted as the de facto standard in platforms like JADE, SPADE |
Frequently Asked Questions
Agent Communication Language (ACL) is the formal protocol enabling autonomous software agents to exchange information, make requests, and coordinate actions. These questions address its core purpose, standards, and role in modern multi-agent systems.
An Agent Communication Language (ACL) is a standardized formal language that defines the syntax, semantics, and pragmatics of messages exchanged between autonomous software agents to enable interoperable knowledge sharing and coordination. It works by providing a structured message format that agents use to make requests, share information, and negotiate. A typical ACL message contains several key components: a performative (the communicative act, like request, inform, or propose), a sender and receiver, a content field carrying the payload (often expressed in a content language like KIF or SL), and a set of conversation control parameters (like a conversation-id and reply-with). Agents interpret these messages based on shared protocols and ontologies, triggering internal reasoning and generating appropriate responses to collaborate on complex tasks.
Enabling Efficiency, Speed & Accuracy
Intelligent Analysis, Decision & Execution
We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.
Talk to Us
Search across company data
Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.
Useful when people spend too long searching or get different answers from different systems.

Automate internal workflows
Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.
Useful when repetitive work moves across multiple tools and teams.

Add AI to products and internal tools
Build assistants, guided actions, or decision support into the software your team or customers already use.
Useful when AI needs to be part of the product, not a separate tool.
Related Terms
Agent Communication Languages (ACLs) are a core component of multi-agent systems. The following terms define the surrounding protocols, patterns, and infrastructure required for agents to interoperate effectively.
Agent Ontology
An ontology provides the shared vocabulary and conceptual model that gives meaning to the content of ACL messages. Without a common ontology, agents may exchange syntactically valid messages with semantically ambiguous or conflicting content. It defines:
- Concepts and their hierarchical relationships.
- Properties and attributes of concepts.
- Formal axioms and constraints. This shared understanding is critical for interoperable reasoning and collaborative task execution.
Agent Communication Protocol
While an ACL defines what can be said, a communication protocol defines the rules of conversation—the allowed sequences of messages. Protocols manage interaction patterns such as:
- Request-Response: A simple query-answer flow.
- Contract Net: A bidding protocol for task allocation.
- Auction Protocols: For resource allocation (e.g., English, Dutch).
- Negotiation Protocols: Structured sequences of offers and counter-offers. Protocols ensure interactions are orderly, predictable, and can reach termination.
Agent Interoperability
This is the primary goal of standardized ACLs: enabling agents built on different platforms, by different teams, or for different purposes to work together. True interoperability requires alignment on multiple layers:
- Transport Layer: How messages are physically delivered (e.g., HTTP, gRPC).
- Syntax Layer: The message format (e.g., FIPA ACL string, JSON-LD).
- Semantic Layer: Shared ontologies for content meaning.
- Pragmatic Layer: Agreed-upon protocols and interaction patterns.
Speech Act Theory
This linguistic and philosophical theory is the conceptual foundation for most ACLs. It posits that utterances are actions (speech acts) with three components:
- Locutionary Act: The literal utterance.
- Illocutionary Act: The intent or force (e.g., promising, requesting).
- Perlocutionary Act: The effect on the listener.
ACLs operationalize this by encoding the illocutionary force as a performative (e.g.,
inform,request), making agent communication a form of action in a social context.

About the author
Prasad Kumkar
CEO & MD, Inference Systems
Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.
His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.
Partnered with leading AI, data, and software stack.
How We Work
Custom AI workflows for your Business
One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.
01
Review the use case
We understand the task, the users, and where AI can actually help.
Read more02
Pick the right approach
We define what needs search, automation, or product integration.
Read more03
Build the first useful version
We implement the part that proves the value first.
Read more04
Improve from there
We add the checks and visibility needed to keep it useful.
Read moreThe first call is a practical review of your use case and the right next step.
Talk to Us