Build your own AI Workbench by bringing an AI agent and configuring it according to the guides in this section. Once set up, use it to write TQL pipelines, understand OCSF schemas, generate parsers, and create data mappings.
Two ways to get AI assistance
Section titled “Two ways to get AI assistance”Tenzir offers two integration points for AI agents:
- MCP Server: A Model Context Protocol server that exposes Tenzir functionality as tools. Works with any MCP-compatible agent.
- Claude Plugins: Skills, subagents, slash commands, and hooks that integrate deeply with Claude Code.
Which should you use?
Section titled “Which should you use?”Claude Code users: install our plugins. We maintain plugins that encode how we write TQL, document features, and manage releases. Skills auto-activate based on context, subagents handle specialized tasks, slash commands provide quick actions, and hooks automate workflows. This is the recommended experience.
Other agents with skills support: The skills specification is open, but you’d need to port our plugins to your agent’s format. We currently only ship plugins for Claude Code.
All other agents: use the MCP server. If your agent doesn’t support skills,
the MCP server gives you access to core tools like make_parser,
make_ocsf_mapping, and run_pipeline. Any MCP-compatible agent works with the
server.
Getting started
Section titled “Getting started”- Install the MCP server: Connect your AI agent to Tenzir.
- Configure your agent: Optimize settings for Tenzir development.
- Use Claude plugins: Install the TQL, OCSF, and Docs plugins for enhanced workflows.
- Generate a parser: Create TQL parsers from sample log events.
- Generate an OCSF mapping: Map parsed data to the OCSF schema.