Guides
Trace backend failures from API requests through server logs and database state to verified fixes: request → logs → data → verify. Use MCP when a coding agent needs controlled runtime evidence for the investigation.
Choose the problem that matches your current investigation. The guides cover investigation workflows and supporting habits, with links to the relevant features and documentation.
Guide 1
How can MCP help troubleshoot a backend issue safely?
Use an evidence-first loop across API requests, server state, and database state while keeping scope and high-risk actions under user control.
Read the MCP troubleshooting guideGuide 2
How do you trace an API error through the server and database?
Start from a reproducible request, follow its evidence into the runtime and persisted state, then verify the result with the original request.
Read the API error debugging guideGuide 3
Why save debugging context before a backend failure?
Save requests and connections as they become useful during development, so later troubleshooting has somewhere to start.
Read the debugging context guideGuide 4
What does local-first mean for backend development?
Separate local workspace data and credentials from the remote systems you access and the optional cloud services you choose to enable.
Read the local-first development guideNext steps
Check what a tool can do, configure it for your project, or follow a troubleshooting workflow.
Understand what a released Unfour capability does, its boundaries, and why it matters.
Explore featuresInstall, configure, and use the released API, SSH, Database, and MCP capabilities.
Read documentationChoose a workflow for a concrete symptom, decision, trade-off, or troubleshooting problem.
Browse guidesDownload Unfour to apply the workflow, or open the documentation for product setup and operating details.
