JarvisCore OSS
FreeApache 2.0, self-managedAutoAgent, CustomAgent, Mesh, memory and HITL with community support; you supply the LLM provider. Official source 1 ↗
Website-checked tool profile
JarvisCore
Open-source Python framework for multi-agent systems that run unattended, with memory, tracing and human review.
Sources checked
- Pricing
- The framework is free under Apache 2.0, including commercial use with no revenue or user cap; you pay your own LLM provider. JarvisCore Enterprise is a commercial agreement for extra modules, managed deployment and support, with no public price. Official source 1 ↗
- Available as
- Python 3.10+ package installed with pip; a minimal agent needs only an LLM API key, and Docker is optional for local credential and memory services. Official source 1 ↗
- Primary group
- Automation & Agents
What it does
JarvisCore is an open-source Python framework for multi-agent systems built to run unattended for weeks, remember past work, fail visibly and keep a readable record of what they did.
Official source 1 ↗Features
- AutoAgent and CustomAgent. AutoAgent runs a full observe-orient-decide-act loop; CustomAgent exposes the loop for deterministic control on the same infrastructure. Official source 1 ↗
- Peer-to-peer agent mesh. Agents discover and message each other over SWIM gossip and ZMQ, with no central orchestrator, on one process or across machines. Official source 1 ↗
- Four-tier memory. Working scratchpad, episodic ledger, compressed long-term summaries and optional cross-session semantic memory. Official source 1 ↗
- Integrations without raw credentials. Prebuilt actions for Slack, GitHub, Zoom, SAP, NetSuite, Microsoft Graph, Salesforce and more; the Nexus layer manages OAuth2 and API keys outside agent reasoning. Official source 1 ↗
- Tracing and human review. Every agent turn, tool call and LLM request is traced, and decisions below a confidence threshold go to a review inbox. Official source 1 ↗
Pricing & free limits
JarvisCore Enterprise
Contact salesCommercial agreementAdds agent identity and delegated sessions, the Maven cost-efficiency system, self-improving workflow DAGs, managed deployment and commercial support. Official source 1 ↗
Billing details
Billing details were not established by the checked official sources.
Free access & limits
Free access and its limits were not established by the checked official sources.
Best for & limitations
Running a fleet of independent agents across processes that claim durable work by capability, call real business systems and survive restarts. Official source 1 ↗
Moving an existing CrewAI or LangGraph project to a peer-mesh runtime. Official source 1 ↗
- Developer framework. It is a Python library for engineers, configured through code and .env files; there is no visual builder. Official source 1 ↗
- Bring your own model. Hosted inference is only for organisations Prescott Data has issued a promotional token; everyone else supplies a provider key. Official source 1 ↗
- Self-hosting is not fully private. Model and API calls still go to whatever providers you configure. Official source 1 ↗
Similar tools
Tools like JarvisCore
Published tools that share its categories.
LloyalTypeScript platform for AI apps with a local model inside them, from one command to a shippable Mac app.
Visit 
SayonicMac voice assistant that opens apps, types, finds files and adds events from plain spoken requests.
Visit 
n8nVisual AI agents and workflows with model integrations, code steps and human approval controls.
Visit 
BardeenAI lead sourcing, web research, qualification and contact enrichment for go-to-market workflows.
Visit Harness RouterAn open-source tool-selection layer for AI agents, with MCP routing and optional local search.
Visit 
WritesonicAI search visibility research, content generation and agents for SEO and GEO improvements.
Visit