Conduit
Conduit
Docsllms.txtHostingGitHubIntroduction

Getting Started

OverviewInstall ConduitMCP SetupYour First AppStart with AI

Learn

ArchitectureClient vs Admin APIConfiguration

Modules

OverviewAuthenticationAuthorizationDatabaseStorageCommunicationsChatRouterFunctions

Guides

Next.js IntegrationReBAC Team ScopingGitOps State Export

Deployment

Deployment OverviewDocker ComposeKubernetes and HelmLocal from SourceContainer Images

Reference

CLI ReferenceClient APIAdmin APIEnvironment VariablesMCP Tools

Resources

Migration v0.16 → v0.17Legacy DocumentationChangelogFAQGlossaryContributing

Start with AI

Give your LLM a copy-paste prompt to build on Conduit v0.17.

Use this page when you want an AI assistant (Cursor, Claude, ChatGPT, etc.) to build on Conduit without pasting docs manually.

Copy the prompt

Cursor users

Install the Conduit Cursor plugin for skills, rules, and subagents that enforce Client vs Admin API boundaries.

What the prompt enforces

  1. llms.txt — machine-readable doc index
  2. MCP for provisioning — schemas and config at dev time
  3. Client API for runtime — user bearer tokens only
  4. Custom endpoints — no client-side filtering of full collections
  5. Storage proxy — no presigned URLs in browsers

For AI agents

Agents should call list_modules after MCP connect and read Client vs Admin API before writing app code.

Next steps

  • Install
  • MCP setup
  • First app

Your First App

Register a user and query the Client API with curl.

Architecture

Core, router, modules, Redis, and gRPC service discovery.

On this page

Copy the promptCursor usersWhat the prompt enforces