Docs

Insforge

Connect your Insforge project to Agenties so the orchestrator can inspect tables, list storage buckets, and run SQL as part of its normal workflow — tracked alongside code changes, issues, and audit logs.

How to connect

1.Open Settings for the project.
2.Go to Integrations → Insforge.
3.Enter your Insforge Project URL (e.g. https://your-ref.eu-central.insforge.app).
4.Enter your anon key.
5.Save — credentials are stored in your OS keychain.
Note:Credentials are stored in your OS keychain and used only from your local machine. They are never sent to the Agenties cloud backend.

What this integration means

ScopeMeaning
Your project integrationAgenties connects to the Insforge project that you configure — not any Agenties-internal service.
MCP tool accessAgents can list tables, list buckets, and run SQL through the configured Insforge API.
Same orchestration modelInsforge work follows Agenties permissions, audit trail, issue context, and continuity flow.
Scoped to your configurationOnly the project you set in Settings is accessible. Nothing else.

MCP tools

ToolParametersDescription
insforge_list_tables-List tables available in the configured Insforge project.
insforge_list_buckets-List storage buckets available in the configured Insforge project.
insforge_run_sqlsql, params?Execute raw SQL on the configured Insforge project and return rows, row count, and fields.
Warning:insforge_run_sql can read or modify real project data. Start with narrow read-only queries, explain intent before destructive operations, and keep mutations small enough to review.

Recommended workflow

StepWhat the agent should do
1 - InspectCall insforge_list_tables and, when needed, insforge_list_buckets.
2 - Read firstUse insforge_run_sql for narrow SELECT queries before planning changes.
3 - Explain intentState which records, tables, or buckets will be affected and why.
4 - Execute carefullyRun writes only after explicit approval or an accepted issue plan.

Common use cases

Use caseExample
Schema discoveryUnderstand available tables before wiring a feature to backend data.
Data debuggingInspect failed records, unexpected state, or broken user flows.
Storage checksVerify bucket availability before adding upload or asset workflows.
Operational maintenanceRun small, reviewed SQL fixes as part of a documented task.