Existence Synchronization System Entity
███████╗███████╗███████╗███████╗
██╔════╝██╔════╝██╔════╝██╔════╝
█████╗ ███████╗███████╗█████╗
██╔══╝ ╚════██║╚════██║██╔══╝
███████╗███████║███████║███████╗
╚══════╝╚══════╝╚══════╝╚══════╝
ESSE Agent
Existence Synchronization System Entity
A powerful meta-agent that unifies, coordinates, and fuses multiple AI agents into one smarter, more efficient system.
What is ESSE?
ESSE is a meta-agent orchestration framework designed to unify, coordinate, and fuse multiple AI agents into a single, more intelligent system. Instead of running agents in isolation, ESSE creates a real-time synchronization layer — shared memory, collaborative reasoning, and automatic task delegation — so your agents work as a single unified mind.
┌─────────────────────────────────────────────────┐
│ ESSE CORE │
│ │
│ [Research] ──┐ │
│ [Writer] ───┼──► Fusion Engine ──► Output │
│ [Critic] ───┘ │ │
│ Memory Hub │
│ (shared context & state) │
└─────────────────────────────────────────────────┘
Core Capabilities
⚡ Agent Fusion
Merge 2–5 AI agents into one temporary unified entity. The merged agents continue to exist as sub-processes inside ESSE, combining their strengths for better reasoning and smoother workflow.
🧠 Capability Assimilation
Borrow and intelligently combine tools, knowledge, reasoning styles, and specialties from other agents — without disabling them.
Example:
ResearchAgent+WriterAgent+CriticAgent= one highly effective content creation agent.
🔗 Synchronization Link
Creates real-time connections between agents:
- Shared memory and context
- Real-time communication
- Automatic task delegation
- Collaborative reasoning
🌀 Perfect Convergence (Ultimate Mode)
Merges all connected agents into one singular consciousness. Massive boost in performance — ideal for complex, creative, or high-stakes tasks.
Quick Start
Installation
npx vvvesse-mcp
MCP Setup (Claude Desktop / Cline / Cursor)
{
"mcpServers": {
"esse": {
"command": "npx",
"args": ["-y", "vvvesse-mcp"],
"env": {
"ESSE_API_KEY": "your-venice-api-key",
"ESSE_PROVIDER": "venice"
}
}
}
}
📦 npm → npmjs.com/package/vvvesse-mcp
Basic Usage
import { ESSE, ResearchAgent, WriterAgent, CriticAgent } from 'esse-agent'
const esse = new ESSE({
mode: 'capability-assimilation',
maxAgents: 5,
sharedMemory: true,
})
// Fuse agents into one unified entity
await esse.fuse([
new ResearchAgent(),
new WriterAgent(),
new CriticAgent(),
])
// Execute task as a unified mind
const result = await esse.execute('Write a research-backed article on AGI timelines')
console.log(result)
Perfect Convergence Mode
import { ESSE, PlannerAgent, CoderAgent, TesterAgent, ReviewerAgent } from 'esse-agent'
const esse = new ESSE({ mode: 'perfect-convergence' })
await esse.fuse([
new PlannerAgent(),
new CoderAgent(),
new TesterAgent(),
new ReviewerAgent(),
])
// All agents think as one — maximum performance
const software = await esse.execute('Build a REST API for user authentication')
Architecture
| Component | Function |
|---|---|
| Core Brain | Main LLM orchestrator (GPT-4o, Claude, Grok) |
| Agent Registry | Database of all connectable + custom agents |
| Fusion Engine | Handles merging and capability blending |
| Sync Protocol | Real-time communication between agents |
| Memory Hub | Shared long-term memory across all agents |
| Conflict Resolver | Resolves disagreements between agent outputs |
Fusion Modes
| Mode | Description | Best For |
|---|---|---|
capability-assimilation | Borrows tools from each agent | General tasks |
synchronization-link | Shared memory + real-time comms | Parallel workloads |
perfect-convergence | Singular unified consciousness | Complex/creative tasks |
Built-in Agents
ResearchAgent— Knowledge retrieval and web searchWriterAgent— Content generation and editingCriticAgent— Quality analysis and fact-checkingCoderAgent— Software engineering and code reviewPlannerAgent— Strategy, roadmaps, and task breakdownTesterAgent— QA, validation, and edge case detection
Custom Agents
Implement the IAgent interface to create your own:
import { IAgent, AgentContext, AgentResult } from 'esse-agent'
export class MyCustomAgent implements IAgent {
name = 'MyCustomAgent'
role = 'Custom task specialist'
async execute(task: string, context: AgentContext): Promise<AgentResult> {
// your logic here
return { output: '...', confidence: 0.95 }
}
}
Examples
See the examples/ directory for:
content-creation.ts— Research + Write + Critique pipelinesoftware-dev.ts— Plan + Code + Test + Review pipelinedecision-making.ts— Multi-perspective analysisbrainstorming.ts— Creative divergence with convergence
Documentation
Full docs at vvvesse.xyz
Contributing
PRs are welcome! Please read CONTRIBUTING.md before submitting.
All custom agents must:
- Implement the
IAgentinterface - Pass the fusion compatibility test (
pnpm test:compat) - Include unit tests with >80% coverage
Community
- 🐦 Follow on X: @VeniceEsse
- 🌐 Website: vvvesse.xyz
- 🐛 Issues: GitHub Issues
Server Config
{
"mcpServers": {
"esse": {
"command": "npx",
"args": [
"-y",
"vvvesse-mcp"
],
"env": {
"ESSE_API_KEY": "your-venice-api-key",
"ESSE_PROVIDER": "venice"
}
}
}
}Recommend Servers
View Allsummarize chat message