Expand description
PulseHive — Shared Consciousness SDK for Multi-Agent AI Systems.
This is the meta-crate that re-exports pulsehive-core and pulsehive-runtime.
Use feature flags openai and anthropic to include LLM providers.
§Quick Start
[dependencies]
pulsehive = { version = "2.0", features = ["openai"] }ⓘ
use pulsehive::prelude::*;
use pulsehive::HiveMind;
let hive = HiveMind::builder()
.substrate_path("my_project.db")
.llm_provider("openai", my_provider)
.build()?;Modules§
- agent
- Agent definition types and workflow composition.
- approval
- Human-in-the-loop approval primitives.
- context
- Context budget configuration for the Perceive phase.
- embedding
- Embedding provider abstraction for domain-specific embedding models.
- error
- Error types for PulseHive SDK.
- event
- Event system for real-time observability into agent execution.
- export
- Event export trait for streaming HiveEvents to external observability systems.
- lens
- Lens-based perception for agent substrate access.
- llm
- LLM provider abstraction and message types.
- prelude
- Re-exports of the most commonly used types.
- tool
- Tool trait and execution context.
Structs§
- Default
Experience Extractor - Default experience extractor using simple rule-based logic.
- Hive
Mind - The central orchestrator of PulseHive.
- Hive
Mind Builder - Builder for constructing a
HiveMindwith validated configuration. - Task
- A task to be executed by deployed agents.