MCP Server
AILinux TriForce MCP
Connect MCP-compatible AI assistants directly to the AILinux tool ecosystem. One endpoint provides model routing, search, code and workspace tools, memory, device integration, automation, federation and more — with capability-scoped access and policy checks.
https://api.ailinux.me/v1/mcpOne MCP connection. A whole capability layer.
MCP lets an AI use tools instead of only generating text. TriForce acts as the AILinux control plane: it exposes a canonical tool registry, routes calls to the correct backend or connected workspace, and applies access policy at execution time.
AI & Models
List models, route prompts across supported providers, use specialist routing and connect model workflows to the rest of TriForce.
Search & Research
Unified search for websites, documentation, code, files, downloads, news, science, images and videos.
Code & Workspace
Inspect trees, read and search code, edit files, use Git and expose explicitly shared local workspaces through AILinux Helper.
Memory
Search persistent memory and, where permissions allow, store structured facts, decisions, code notes, summaries and tasks.
Devices & Helper
Pair desktop or Android endpoints and expose only the capabilities that were explicitly granted to the MCP session.
Automation & Services
Work with notifications, services, containers, agents and other TriForce integrations through policy-controlled tools.
Federation
Target local or remote nodes through the same capability model instead of maintaining separate tool implementations.
Progressive Discovery
Tool metadata includes inventory, task group, scope, access hints and execution target so clients can load only what they need.
Use the canonical endpoint in any compatible MCP client.
For clients that accept a remote MCP server URL, use the following endpoint. Authentication and available tools depend on the client identity, account and attached capability grants.
Server name: AILinux TriForce
Transport: Streamable HTTP
URL: https://api.ailinux.me/v1/mcpDiscovery is not unrestricted execution.
A visible tool does not automatically mean an anonymous client can execute it. Privileged actions remain protected by authentication, RBAC, policy, workspace leases and per-capability grants.
Public guest profile
Credential-less MCP clients can use the safe public surface. Sensitive persistent or privileged actions remain restricted.
Scoped local access
Local files, code and device functions only become available after explicit workspace/device pairing and capability sharing.
Call-time authorization
Tool visibility helps discovery; authorization is still checked when a tool is actually called.
Current MCP tools
This list is loaded directly from the production MCP server. When the canonical TriForce registry changes, this page updates automatically.
Connect once. Discover capabilities dynamically.
Use the canonical TriForce MCP endpoint in your compatible AI client, or open the server directly to inspect the current service.