Quickstart¶

Get Neuralgentics from "Clone" to "First Dispatch" in under 5 minutes.

🚀 5-Minute Bootstrap¶

1. Install the Runtime¶

If you haven't already, run the binary installer:

curl -fsSL https://raw.githubusercontent.com/Veedubin/neuralgentics/main/scripts/install.sh | bash

2. Initialize Infrastructure¶

Neuralgentics requires a PostgreSQL instance with pgvector and the Memini-AI server.

# Start the development database and memory server
./scripts/dev-up.sh

3. Configure Your LLM¶

Edit your .env file or export the required keys:

export OLLAMA_API_KEY="your_key_here"
export NEURALGENTICS_API_KEY="your_key_here"

4. Launch the TUI¶

neuralgentics

5. Your First Dispatch¶

Once inside the TUI, try a simple orchestrator task: "Analyze the current project structure and create a high-level map of the Go modules."

What happens under the hood: 1. Orchestrator queries memory for project context. 2. Architect is dispatched to design the map. 3. Explorer is dispatched to find the files. 4. Writer is dispatched to format the final report.

✅ Verification Checklist¶

Step	Action	Expected Result
1	`neuralgentics status`	All components (Backend, Broker, Memory) show `ONLINE`
2	`neuralgentics --version`	Returns `v0.6.5` (or whatever you installed)
3	First Task	Task transitions from `triage` $\rightarrow$ `running` $\rightarrow$ `done`

If you encounter any issues, check the Troubleshooting Guide.