Scout on-premise setup

Deploy Scout AI agent in your Kubernetes cluster with full control over data processing

Prerequisites

Before deploying Scout on-premise, you need:

• Kubernetes cluster with kubectl access
• OpenAI-compatible API (e.g. LiteLLM) serving an Anthropic Claude model
• Coalesce Quality OAuth credentials from on-premise integration setup:
  • SYNQ_CLIENT_ID
  • SYNQ_CLIENT_SECRET

Deployment

Scout runs as a single process that only ever connects outward, so it can sit inside your network with nothing exposed. Two supported shapes:

Kubernetes

The recommended shape for a long-running deployment. All configurations, detailed instructions and examples are in our official repository:

getsynq/synq-scout-k8s

The repository provides complete deployment guidance including:

• Kubernetes configurations
• Environment setup instructions
• Sample deployment files
• Troubleshooting guides

Binary

For evaluating Scout, or for running it on a host you manage rather than a cluster. Pick a version from the releases page — the archive filename carries it, so it has to be named explicitly:

VERSION=0.2.0
OS=$(uname -s | tr '[:upper:]' '[:lower:]')          # darwin or linux
ARCH=$(uname -m | sed 's/x86_64/amd64/; s/aarch64/arm64/')

curl -fL "https://github.com/getsynq/synq-scout/releases/download/v${VERSION}/synq-scout_${VERSION}_${OS}_${ARCH}.tar.gz" \
  | tar -xz
sudo mv synq-scout /usr/local/bin/

Builds are published for macOS and Linux on both amd64 and arm64, and every release ships a checksums.txt you can verify with sha256sum -c checksums.txt --ignore-missing. Confirm which build you have with synq-scout --version.

Each archive carries an agent.example.yaml to start from. Check the whole setup before running anything real:

synq-scout health

Then start the agent:

synq-scout agent

Docker

docker pull europe-docker.pkg.dev/synq-cicd-public/synq-public/synq-scout:latest

The same image the Kubernetes manifests deploy, if you would rather run it under your own container runtime.

You do not need any of this to use Scout’s tools from an AI client — the MCP server is hosted. See MCP.

Configuration

During your on-premise integration setup in Coalesce Quality, you’ll receive OAuth credentials that are required for deployment:

• SYNQ_CLIENT_ID
• SYNQ_CLIENT_SECRET

These credentials authenticate Scout with the Coalesce Quality platform and must be configured in your deployment.

LLM Model Configuration

Scout uses two model roles — a thinking model (for reasoning and triage) and a summary model (for generating summaries). Both default to Claude Sonnet when not explicitly configured.

You can override the models using environment variables:

Variable	Description
SYNQ_THINKING_MODEL	Model used for reasoning and triage
SYNQ_SUMMARY_MODEL	Model used for generating summaries
OPENAI_MODEL	Fallback model for both roles if the above are not set

The model names should match what your OpenAI-compatible API expects (e.g. the model name configured in LiteLLM).

To see the list of currently supported models and their stable/latest versions, go to Settings → Scout AI → Models in the Coalesce Quality platform.

Support

For deployment assistance:

• Repository Documentation - Complete deployment guide
• Repository Issues - Report deployment issues
• Coalesce Quality support - Contact your team for assistance

On-premise deployment provides complete control over your data while maintaining Scout’s full functionality and enterprise security requirements.