Two agent profiles
Choose AutoAgent for a framework-managed observe, orient, decide, and act loop, or CustomAgent for direct, deterministic control over execution.
JarvisCore
JarvisCore is an open-source Python framework for developers building autonomous multi-agent systems. It provides agent profiles, peer-to-peer coordination, persistent memory, credential isolation, observability, and human review workflows.
Open-source framework for autonomous multi-agent systems
JarvisCore is an open-source Python framework for building autonomous multi-agent AI systems. It combines agent execution, peer-to-peer communication, memory, credential handling, human review, and tracing in a shared runtime for agents that may operate without continuous supervision.
The framework offers two operating models: AutoAgent manages the cognitive loop for tasks such as research, analysis, content work, and data extraction, while CustomAgent exposes execution to developers writing plain Python for API workers and deterministic pipelines. A minimal setup requires Python 3.10 or later and one configured LLM provider; additional capabilities are enabled through optional package extras and services.
Core capabilities
Layered memory
Working scratchpad, episodic ledger, compressed long-term summaries, and optional cross-session semantic memory preserve context across restarts. Athena MemOS can be connected through `ATHENA_URL`.
Peer-to-peer coordination
Agents discover and message one another through a PeerClient API over SWIM gossip and ZMQ, allowing nodes to join, fail, and rejoin without a central orchestrator.
Credential isolation
Nexus manages OAuth2, API keys, and basic authentication so third-party credentials remain outside agent reasoning.
Operational observability
Agent turns, tool calls, and LLM requests are traced automatically, with Redis PubSub live streams, JSONL output for compliance workflows, and an optional Prometheus metrics endpoint.
Human review checkpoints
HITLQueue can pause decisions that exceed configured confidence thresholds, send them to a review inbox, and resume execution after a human response.
Practical use cases
Autonomous research and analysis
Use AutoAgent to conduct research and analysis tasks with a role, capability list, and system prompt, while retaining context across sessions when memory services are configured.
Deterministic API workers
Use CustomAgent to wrap existing APIs or services in explicit Python execution logic when predictable control matters more than framework-managed reasoning.
Distributed agent workflows
Run several specialized agents as a research or operations network. Peer discovery and messaging allow agents to coordinate across one process or distributed machines.
Tool-connected business processes
Connect service actions and third-party APIs through the integration catalog while keeping OAuth2, API keys, and basic authentication in the Nexus credential layer.
Human-supervised automation
Add review checkpoints to workflows where selected decisions need human approval, then resume the agent after a response is submitted.
Pros and Cons
Pros
- Supports both framework-managed autonomous agents and developer-controlled deterministic agents.
- Provides peer-to-peer coordination for single-process or distributed deployments rather than requiring a central orchestrator.
- Combines durable memory, credential isolation, observability, and human-in-the-loop review in the same framework.
- Can start with a minimal agent using only an LLM API key, with Docker optional for the basic workflow.
Cons
- Advanced capabilities may require optional Python extras, external services, or local infrastructure such as Redis, Athena, Nexus, or a sandbox.
- The framework is Python-based and requires Python 3.10 or later.
- The available product materials do not establish pricing, plan limits, deployment guarantees, or team-support terms.
FAQ
What is required to get started with JarvisCore?
JarvisCore requires Python 3.10 or later. A minimal agent can run with only an LLM API key; Docker is optional unless you use local infrastructure or a containerized sandbox.
How do I create a first agent?
Install the package with pip, run `jarviscore init`, copy `.env.example` to `.env`, configure one supported LLM provider, and run `jarviscore check`. The site provides a first-agent example using `Mesh` and `AutoAgent`.
What is the difference between AutoAgent and CustomAgent?
AutoAgent provides the framework's full cognitive loop, including LLM reasoning, sandboxed code generation, and self-repair. CustomAgent exposes the execution loop so developers can implement deterministic Python logic directly.
Can JarvisCore run distributed multi-agent systems?
The framework can run agents in one process or across distributed machines. Distributed deployments use the SWIM gossip protocol and ZMQ transport, while Redis can support distributed workflows, cross-session state, and peer routing when configured.
Quick Facts
- Category
- Developer tool / AI agent framework
- License model
- Open source
- Primary language
- Python
- Minimum runtime
- Python 3.10 or later
- Agent profiles
- AutoAgent and CustomAgent
- Basic setup
- One LLM provider key; Docker is optional for a minimal agent
Alternativas a JarvisCore
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
GitBot
GitBot is a local browser-based workspace for creating reusable coding bots powered by Claude Code, Codex, or OpenCode. It helps developers run those agents across repository folders with persistent threads and configurable permissions.
AakarDev AI
AakarDev AI helps teams manage AI provider access, project-level setups, logs, and analytics from one dashboard. It supports BYOK workflows and lists providers including OpenAI, Google Gemini, Anthropic, Groq, Mistral AI, and Perplexity AI.
Trigger.dev chat agent
Trigger.dev chat agent is a durable AI chat backend for developers building stateful conversations that can survive refreshes, crashes, and long-running turns. It connects with the AI SDK `useChat` flow and runs on managed infrastructure with no timeout on a turn.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
Codex Plugins
Codex Plugins bundle reusable skills, app integrations, and MCP servers into workflows you can install in the Codex app or use from Codex CLI. They help extend Codex with connected-service tasks, reusable instructions, and shared team workflows.