An AI-powered hub that connects disparate enterprise systems (ERP, CRM, databases) via LLM agents with dynamic DAG planning and concurrent execution.
FIM One is an LLM-powered agent runtime that connects and orchestrates disparate enterprise systems like ERP, CRM, and databases. It solves the problem of siloed business systems by providing an AI-powered hub that enables dynamic workflow planning and concurrent execution without modifying existing infrastructure.
Developers and IT teams in enterprises needing to integrate legacy systems, automate cross-system workflows, or embed AI copilots into existing applications.
Developers choose FIM One for its flexible deployment modes (Standalone, Copilot, Hub), dynamic DAG planning that eliminates hard-coded workflows, and its production-ready agent harness with features like error recovery and multi-tenant support.
Open-source agent platform for Global × China enterprises — wire every system through one agent core. Self-hosted, any LLM.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Unifies APIs, databases, and MCP servers into a single interface with auto-registered agent tools and auth injection, enabling connections to systems like SAP without infrastructure changes, as highlighted in the Connector Hub feature.
LLM decomposes goals into dependency graphs at runtime, allowing parallel execution of independent steps via asyncio and eliminating the need for hard-coded workflows, a core feature from the Planning & Execution section.
Includes a ReAct loop with automatic error recovery, configurable auto-routing, and Hook middleware for deterministic guardrails, providing a robust execution environment as described in the agent harness details.
Features JWT authentication, organization isolation, admin analytics, and a marketplace for publishing agents and connectors, making it suitable for scalable enterprise deployments, per the Multi-Tenant Platform section.
The Source Available License is not OSI-approved and prohibits multi-tenant SaaS, competing platforms, and white-labeling, limiting commercial and derivative use cases, as stated in the License section.
Requires Docker or local installation with multiple prerequisites like Python 3.11+, Node.js, uv, and pnpm, which can be a barrier for quick adoption, as shown in the Quick Start and Local Development steps.
Heavily relies on OpenAI-compatible LLMs and specific services like Jina AI for embeddings, creating potential lock-in and ongoing cost concerns, evident from the Configuration and RAG pipeline dependencies.