RadOps
VerifiedMulti-agent platform for autonomous DevOps automation with built-in verification.
What is RadOps?
RadOps is an open-source multi-agent framework designed to automate DevOps operations. It replaces simple chat interfaces with agents that maintain context across steps, follow defined procedures, and check their own results before reporting back.
The platform organizes work through a supervisor-worker model, three levels of memory, and skills defined in Markdown. Agents discover each other via prompts or vector search, connect to tools through the Model Context Protocol, and pause for human approval on sensitive actions.
It targets DevOps teams and platform engineers who need reliable automation for infrastructure checks, workflow execution, and knowledge retrieval without writing new code for each specialist agent.
What you can build with RadOps
Host Health Checks
Decompose requests into DNS lookup, HTTP status, and multi-node ping steps, then produce a verified summary report.
Standard Operating Procedures
Convert Markdown-defined SOPs into executable sequences that call tools, skills, and external services with variable substitution.
Specialized Agent Deployment
Create new agents for network or security tasks by editing YAML persona and tool definitions without additional coding.
Install RadOps
git clone https://github.com/mehrdadrad/radops.git
cd radops- 1Clone the repository from GitHub and ensure Python 3.11 is installed.
- 2Install required dependencies listed in the project files.
- 3Configure LLM provider credentials and vector database settings in the appropriate config files.
- 4Define initial agents or workflows using the provided YAML and Markdown templates.
- 5Launch the platform and test with a sample workflow such as an infrastructure health check.
RadOps: pros & cons
Pros
- +Strong separation of memory types and built-in QA auditing reduce hallucinations
- +Config-driven agents and skills allow rapid extension without code changes
- +Supports both small-team prompt mode and large-scale vector discovery
- +Human-in-the-loop controls and OpenTelemetry tracing improve operational safety
Cons
- –Setup requires familiarity with multiple LLM and database integrations
- –No pre-built package; users must clone and configure from source
- –Complex workflows still need careful Markdown and YAML authoring
Frequently asked questions
It works with OpenAI, Anthropic, Azure, Google, Groq, Mistral, AWS Bedrock, DeepSeek, and local Ollama models that support tool calling.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar agents
Other general-purpose options worth comparing.
