Skip to main content

Crate modular_agent_core

Crate modular_agent_core 

Source
Expand description

Modular Agent Core - A crate for building modular multi-agents intelligent systems.

This crate provides a set of tools and abstractions to create, configure, and run agents in a stream-based architecture. It includes support for defining agent behaviors, managing agent flows, handling agent input and output.

Re-exports§

pub extern crate self as modular_agent_core;
pub use llm::Message;
pub use llm::ToolCall;
pub use llm::ToolCallFunction;
pub use photon_rs;
pub use im;
pub use inventory;
pub use fnv;
pub use indexmap;

Modules§

llm
mcp
tool

Structs§

AgentConfigSpec
AgentConfigs
AgentContext
Event-scoped context that identifies a single flow across agents and carries auxiliary metadata.
AgentData
The core data structure for an agent.
AgentDefinition
AgentRegistration
Registration entry emitted by the #[modular_agent] macro.
AgentSpec
Information held by each agent.
ConnectionSpec
ModularAgent
PhotonImage
Provides the image’s height, width, and contains the image’s raw pixels. For use when communicating between JS and WASM, and also natively.
Preset
PresetInfo
PresetSpec

Enums§

AgentError
AgentStatus
AgentValue
MAKEvent

Traits§

Agent
The core trait for all agents.
AgentOutput
AsAgent
HasAgentData

Functions§

agent_new
new_agent_boxed

Type Aliases§

AgentConfigSpecs
AgentConfigsMap
AgentDefinitions
AgentValueMap
FnvIndexMap
FnvIndexSet
PresetSpecs

Attribute Macros§

async_trait
modular_agent
Declare agent metadata and generate agent_definition / register helpers.