Autonomous AI Platform

CH8 Agent
Intelligent AI Orchestration

Autonomous AI agents that integrate with anything—databases, APIs, cloud storage, MCP servers. Works standalone, in clusters, or distributed across any hardware. 40-60% token savings through recurrent learning. Infinite memory. 100% local and free.

Get Started GitHub
3
Operation Modes
40-60%
Token Savings
2-3x
Faster
∞
Memory

Autonomous AI Agents

Self-organizing agents that integrate with databases, APIs, storage, and MCP servers. Zero configuration needed.

🔍

Data Extraction

10 specialized agents for XML, JSON, CSV, PDF, Excel, Parquet, YAML, SQL

Auto-configure
🔗

Integration (MCP)

Connect APIs, databases, file systems, RAG automatically

Zero-setup
🗄️

Database Agents

PostgreSQL, MySQL, MongoDB, Redis with smart querying

Async-ready
☁️

Storage Agents

MinIO, S3, GCS, Azure - Unified multi-cloud interface

Multi-cloud
🧠

LLM Orchestrator

Coordinate multiple small models for complex reasoning

Token-efficient
🎯

Custom Agents

Create your own with Python decorators - plug & play

Extensible

3 Operation Modes

Choose the best mode for your needs—standalone, cluster, or distributed

Standalone Mode

Single device, full power. Run small or large models locally with intelligent task decomposition.

  • Perfect for personal projects
  • Works offline
  • Small models (0.5-1B) or large (7B+)
  • Recurrent learning saves tokens

Cluster Mode

Multiple models on one device working in parallel. Specialized agents for different tasks.

  • 2-3x faster execution
  • Parallel task processing
  • Smart result aggregation
  • Ideal for large integrations

Distributed Mode

Innovative feature: spread work across multiple devices—Raspberry Pi, old laptops, Android phones.

  • Use any hardware you have
  • Democratic AI approach
  • Zero infrastructure cost
  • Runs on devices from 2000s

Enterprise Features

Production-ready capabilities for real-world applications

Recurrent Learning

Agents learn from previous tasks, reducing token usage by 40-60% over time.

  • Infinite memory across sessions
  • Pattern recognition
  • Adaptive task optimization

MCP Integration

Model Context Protocol support for seamless integration with any tool or API.

  • Auto-discovery of services
  • Zero-configuration setup
  • Extensible architecture

Universal Database

Connect to any database—SQL, NoSQL, or object storage—with unified interface.

  • PostgreSQL, MySQL, MongoDB, Redis
  • S3, MinIO, GCS, Azure
  • Async operations

Data Extraction

10+ specialized agents for extracting data from any format.

  • XML, JSON, CSV, Excel, Parquet
  • PDF, YAML, TOML, SQL
  • XPath & column projection

Universal Platform

Runs on any hardware—from Raspberry Pi to GPU servers.

  • Linux, Windows, macOS
  • Raspberry Pi (all models)
  • Android (Termux or native APK)

100% Private

All processing stays local. No cloud dependencies.

  • Fully offline capable
  • No data tracking
  • Zero external API calls

Why CH8 Agent?

Better than traditional approaches and cloud-based solutions

❌

Traditional Approach

  • Large expensive models (7B-70B)
  • Requires powerful GPU servers
  • Slow sequential processing
  • High cost per request
  • Cloud dependency
  • No memory between sessions
VS
✅

CH8 Agent

  • 3 modes: Standalone, Cluster, Distributed
  • Small (0.5-1B) AND large models
  • Runs on ANY hardware
  • 40-60% token savings
  • Infinite memory (recurrent learning)
  • MCP servers integration
  • Zero cost (use what you own)
  • 100% local & private

Real World Example

The "Drawer Cluster" - Using old devices found in drawers

💻

Old Thinkpad (2011)

4GB RAM, Linux 32-bit

Phi-3 Mini Q4 Reasoning
🥧

Raspberry Pi 3

1GB RAM

TinyLlama Q2 Extraction
🖥️

Mac Mini 2012

8GB RAM

Gemma-2B Q4 Aggregation
📱

Android Phone

2GB RAM

Cloud (Groq) Fast Tasks
Total Cost: $0 (hardware already owned)
Power: ~30W total (runs 24/7)
Performance: 100+ tasks/hour
Speed: 14 seconds (vs 45s single model)

Get Started in 60 Seconds

One-line install for any platform

curl -fsSL https://raw.githubusercontent.com/hudsonrj/ch8-cluster-agent/main/scripts/install.sh | bash
curl -fsSL https://raw.githubusercontent.com/hudsonrj/ch8-cluster-agent/main/scripts/install-macos.sh | bash
powershell -c "iwr -useb https://raw.githubusercontent.com/hudsonrj/ch8-cluster-agent/main/scripts/install.ps1 | iex"
curl -fsSL https://raw.githubusercontent.com/hudsonrj/ch8-cluster-agent/main/scripts/install-rpi.sh | bash
# Install Termux, then:
curl -fsSL https://raw.githubusercontent.com/hudsonrj/ch8-cluster-agent/main/android/install-termux.sh | bash

Start Building with CH8 Agent

Open source, free forever. Deploy in standalone, cluster, or distributed mode.

Install Now GitHub