Skip to main content

P2P distributed agent framework with LLM code generation and event-sourced state management

Project description

JarvisCore Framework

P2P distributed agent framework with LLM code generation and production-grade state management

Features

  • Simple Agent Definition - Write just 3 attributes, framework handles everything
  • P2P Mesh Architecture - Automatic agent discovery and task routing via SWIM protocol
  • Event-Sourced State - Complete audit trail with crash recovery
  • Autonomous Execution - LLM code generation with automatic repair

Installation

pip install jarviscore

Setup & Validation

1. Configure LLM Provider

Copy the example config and add your API key:

cp .env.example .env
# Edit .env and add one of: CLAUDE_API_KEY, AZURE_API_KEY, GEMINI_API_KEY, or LLM_ENDPOINT

2. Validate Installation

# Check setup
python -m jarviscore.cli.check

# Test LLM connectivity
python -m jarviscore.cli.check --validate-llm

# Run smoke test (end-to-end validation)
python -m jarviscore.cli.smoketest

All checks pass? You're ready to build agents!

Quick Start

from jarviscore import Mesh
from jarviscore.profiles import PromptDevAgent

# Define agent (3 lines)
class ScraperAgent(PromptDevAgent):
    role = "scraper"
    capabilities = ["web_scraping"]
    system_prompt = "You are an expert web scraper..."

# Create mesh and run workflow
mesh = Mesh(mode="autonomous")
mesh.add(ScraperAgent)
await mesh.start()

results = await mesh.workflow(
    workflow_id="wf-123",
    steps=[
        {"id": "scrape", "task": "Scrape example.com", "role": "scraper"}
    ]
)

Architecture

JarvisCore is built on three layers:

  1. Execution Layer (20%) - Profile-specific execution (Prompt-Dev, MCP)
  2. Orchestration Layer (60%) - Workflow engine, dependencies, state management
  3. P2P Layer (20%) - Agent discovery, task routing, mesh coordination

Documentation

Development Status

Version: 0.1.0 (Alpha) Day 1: Core framework foundation ✅

License

MIT License - see LICENSE file for details

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

jarviscore_framework-0.1.0.tar.gz (102.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

jarviscore_framework-0.1.0-py3-none-any.whl (121.0 kB view details)

Uploaded Python 3

File details

Details for the file jarviscore_framework-0.1.0.tar.gz.

File metadata

  • Download URL: jarviscore_framework-0.1.0.tar.gz
  • Upload date:
  • Size: 102.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for jarviscore_framework-0.1.0.tar.gz
Algorithm Hash digest
SHA256 b72566e3669a92fb7f1ff381748e2106fe7d6f482d5e914ea7363f92b6841585
MD5 33263bcca6fa21bb08a94c26321daaa5
BLAKE2b-256 07247579b46ef3c471e43acf62f35f0f134fecef14910e5d9a074658c0d3f8ea

See more details on using hashes here.

File details

Details for the file jarviscore_framework-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for jarviscore_framework-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 53d895c7d57479181da4f6863255101f652cc644bb3d4e43244745126ceb95c5
MD5 10b8edce1df8c8222f2afa8f9fc7c8f4
BLAKE2b-256 3de8f6958b7e55f153b7683c09b35115bf48871e911262b9e96fc2095371bc43

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page