Hermes Atlas
Models, providers & proxies · works with Hermes Agent

LinkMind

landingbj/LinkMind

AI middleware for model routing, RAG and OpenAI-compatible APIs, with Hermes Agent sync hooks

In short

LinkMind is an enterprise AI middleware layer that routes between model providers, adds RAG and multimodal APIs, and exposes OpenAI-compatible endpoints. It includes runtime sync hooks for Hermes Agent, OpenClaw, DeerFlow and OpenHuman.

What LinkMind does

LinkMind is multimodal AI middleware that sits between business systems, private knowledge, model providers and agent runtimes. One layer covers chat, RAG, OCR, speech recognition and synthesis, image and video workflows, text-to-SQL, embeddings, rerank, MCP access, skills, worker orchestration and OpenAI-compatible APIs. Multi-model routing, failover and orchestration are configured centrally in lagi.yml, and RAG can connect to Chroma, Elasticsearch, Milvus, MySQL, Pinecone and SQLite.

For agent users, the README says the codebase includes runtime sync hooks for OpenClaw, Hermes Agent, DeerFlow and OpenHuman. When one of these frontends already exists locally, you can connect it as an optional integration and keep configuration synchronized rather than rebuilding workflows. LinkMind runs as one unified runtime that can be deployed as a team AI hub on an appliance, server, personal computer or Docker. It also lists Medusa cache acceleration, token statistics, filters and runtime governance for production cost control.

Key features

  • One middleware layer for chat, RAG, OCR, speech, image and video workflows and text-to-SQL
  • Central multi-model routing and failover in lagi.yml
  • RAG connectors for Chroma, Elasticsearch, Milvus, MySQL, Pinecone and SQLite
  • OpenAI-compatible APIs, MCP access and skills
  • Integration hooks for OpenClaw, Hermes Agent, DeerFlow and OpenHuman
  • Token statistics, filters and Medusa cache acceleration

When to use it

  • Give a team one gateway to many model providers with failover
  • Connect an existing Hermes Agent to a shared model and knowledge layer
  • Put private-knowledge RAG behind OpenAI-compatible endpoints

Who it is for: Teams that need one governed layer between business systems, model providers and local agent runtimes.

How it fits with Hermes Agent

Hermes Agent is listed under local agent frameworks, and the README says LinkMind has an integration hook that can sync configuration with an existing Hermes Agent as an optional connection.

How to install LinkMind

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

curl -fsSL https://cdn.linkmind.top/install.sh | bash
docker pull landingbj/linkmind
docker run -d -p 8080:8080 landingbj/linkmind

Requirements: JDK 8 or later for the installer and the packaged jar

Note: GitHub could not identify the license (NOASSERTION), so reuse terms need to be checked in the repository.

FAQ

What is LinkMind?

LinkMind is enterprise-grade multimodal AI middleware. It sits between business systems, private knowledge, model providers and agent runtimes, and handles routing, failover, RAG and OpenAI-compatible APIs.

Does LinkMind work with Hermes Agent?

Yes, as an optional integration. The README lists Hermes Agent among its local agent frameworks and says LinkMind includes runtime sync hooks for it, alongside OpenClaw, DeerFlow and OpenHuman.

How do I install LinkMind?

On macOS or Linux, run the installer script with curl, or use the Docker image with docker pull landingbj/linkmind. The README also offers a packaged jar and a Maven source build, and the web UI opens at http://localhost:8080.

Similar models for Hermes Agent

All models

Related guides: How to install Hermes Agent