Getting Started¶
In this tutorial, you'll clone the VlinderCLI repository, build it from source, and run your first command.
Time: ~15 minutes
Prerequisites: macOS or Linux, a terminal. All other dependencies are installed in Step 1.
Step 1: Install Prerequisites¶
Follow the Installation guide to install:
- Rust toolchain (via rustup)
- protobuf-compiler
- NATS
- Podman
- Ollama (if your agents use LLMs)
Step 2: Clone and Build¶
bash git clone https://github.com/vlindercli/vlindercli.git cd vlindercli cargo build --release export PATH="$PWD/target/release:$PATH"
Step 3: Bootstrap¶
Create the config directory and a minimal config file. See Installation — Bootstrap for the full config template.
bash mkdir -p ~/.vlinder/{agents,conversations,logs}
Create ~/.vlinder/config.toml with at least [queue] and [state] sections.
Step 4: Start Services¶
You need three terminals:
Terminal 1 — NATS: bash nats-server -js
Terminal 2 — Ollama (if using LLM agents): bash ollama serve
Terminal 3 — Vlinder daemon: bash vlinderd
Step 5: Verify¶
In a fourth terminal:
bash vlinder agent list
You should see an empty list — no agents deployed yet.
Next Steps¶
- Your First Agent — scaffold and run your own agent
- Adding Models — pull and register inference/embedding models
- Configuration — customize logging and providers
- Architecture — understand the component model