More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    MCPDecisions Local Setup

    1

    Install MCPLaya on Windows, macOS, or Linux, verify your Laya models offline, and connect the local MCP server to your agent without overwriting unrelated settings.

    $1

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    MCPDecisions Local Setup

    MCPDecisions Local Setup

    Example session with this skill installed

    Set up MCPLaya on my Windows x64 PC for Claude Code. Use the default CPU setup, verify English and multilingual inference offline, and put the usage skill under C:\Users\Alex\Documents\AgentSkills. Keep my other MCP servers.

    • Read your context and instructions
    • Compiled the mcpdecisions local setup

    Illustrative setup report

    MCPLaya was set up for CPU inference on Windows x64. The client must reload before live MCP checks can run.

    CheckExample result
    SourceReviewed MCPLaya revision 0c8fb86382fe0dd7210bc8f71f0e6dc8d5f1ac33 with Laya 0.3.20
    RuntimeIsolated Python 3.12 in the user account. MCPLaya doctor passed.
    ModelsEnglish and multilingual checkpoints cached. Both passed a local inference check with network access disabled.
    ClientAdded the local stdio entry using the absolute path to the installed interpreter. Backed up the configuration and preserved the other MCP servers.
    Usage skillCopied to C:\Users\Alex\Documents\AgentSkills\mcplaya-decisions. This custom folder may require manual loading.

    After reload

    Restart Claude Code, confirm all eight MCPLaya tools are listed, then run laya_status and an explicit English laya_predict request. Tool discovery and the live prediction are pending until that check.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Run a JEV-style decision engine on your PC with Laya, a local alternative to OpenAI Decisions or JEV.

    Set up MCPLaya, a local stdio MCP server for Laya text classification, on the computer where your agent runs.

    Laya helps your agent make structured decisions from text. Give it a message and a set of options, and it scores which option best fits. For example, your agent can classify an incoming request and use the result to route it to the right workflow. Once the model weights are downloaded, Laya runs inference on your PC, even offline.

    • Install the reviewed MCPLaya revision in an isolated Python 3.12 environment using uv.
    • Download the English and multilingual Laya checkpoints and verify inference with the installed interpreter. Repeat the check against the local cache offline. The typed-decisions checkpoint is optional.
    • Connect the server to your MCP client, back up the existing configuration, and verify tool discovery and a sample prediction when the client can reload.
    • Copy the public mcplaya-decisions usage skill to the directory you choose.

    Initial setup needs local terminal access, internet, and several GB of available disk space. Agensi delivers the installation skill. The server and models run on your computer.

    A cloud terminal cannot install onto your PC. Windows ARM64, 32-bit systems, and Alpine/musl are outside the target platforms. Intel macOS uses legacy dependency constraints. Some MCP clients require a restart or manual enablement.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Requires

    GitHub
    Astral
    Local terminal on the target computer
    Internet for initial downloads
    Several GB of available disk space
    Python

    Permissions

    Terminal / Shell
    Network Access
    Read Files
    Write Files
    Environment Variables

    Allowed Hosts

    astral.sh
    codeload.github.com

    Uses the local shell to install dependencies, download model weights, set cache paths, back up and update the selected client configuration, and copy the public usage skill into the chosen directory.

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    It installs the reviewed MCPLaya server in an isolated Python 3.12 environment, prepares official Laya checkpoints, and configures a local stdio MCP client.

    The skill targets Windows x64, macOS Apple Silicon or Intel, and Linux glibc x64 or ARM64 with compatible dependencies. Your agent needs a local shell and your MCP client must support stdio.

    The setup can provision uv and Python 3.12 in your account. It can reuse a healthy existing setup.

    Initial downloads and cache files can take several GB. The default prepares English and multilingual checkpoints. You can request an English-only setup.

    The ZIP includes setup instructions, platform and client references, and helper scripts. The usage skill is copied from the pinned public MCPLaya source into your chosen folder.

    The skill runs a bounded classification with each selected checkpoint using the installed Python, then repeats the check in offline mode against the cache.

    New in LLM Ops