Skip to main content
Glama
orgenix-ai

Ori for Muse

by orgenix-ai
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Hybrid

Ori for Muse

Build a real, hosted app from Muse Code, Meta's coding agent. It's powered by Ori, the AI software team from Orgenix.

Tell Muse Code what you want built. Ori's team takes it from there: it shapes the idea with you, designs it, builds it, tests it and puts it online. Muse Code keeps you posted, and you can continue in Ori for the full view.

Built for the Meta Global AI Developer Hackathon (Oct 26 – Nov 4, 2026). The hosted endpoint opens on Nov 1, 2026; until then, run it locally against an Ori instance.

Connect Muse Code in two minutes

  1. Install Muse Code and sign in to your Meta Model API account.

  2. Add Ori to Muse Code's settings (~/.config/muse/settings.json):

    {
  "mcpServers": {
    "ori": { "url": "https://ori-for-muse.orgenix.ai/api/mcp" }
  }
}

  3. Start Muse Code and ask for an app (see docs/demo-prompts.md).

Muse Code hands the idea to Ori and opens Ori in your browser with your idea already in the conversation.

  1. To check on your apps from Muse Code, sign in to Ori once:

    muse mcp login ori

    Your browser opens Ori's sign-in and a short "Connect Muse Code to Ori?" card. After that, ask Muse Code "How are my apps doing?".

Related MCP server: Slipway

Tools

Tool

What it does

Meta class

start_app_with_ori

Hands an app idea to Ori and returns a one-time link (30 minutes) that opens Ori with it

Write (nothing is built or charged)

app_status

Lists your apps: stage, what's waiting on you (like confirming the prototype or going live) with a link to do it in Ori, and the live address. Needs sign-in.

Read

Sign-in is standard MCP OAuth: Ori's authorization server (Supabase Auth's OAuth 2.1 server) uses PKCE and dynamic client registration, and this endpoint publishes its protected resource metadata at /.well-known/oauth-protected-resource. Your access token goes to Ori as is; this service stores nothing.

Run it locally

npm install
ORI_BASE_URL=http://localhost:3110 ORI_CONNECTOR_TOKEN=dev npm run dev

Then use http://localhost:8787/mcp as the server URL in Muse Code's settings.

Variable

Meaning

ORI_BASE_URL

The Ori instance to hand ideas to

ORI_CONNECTOR_TOKEN

This service's key for Ori's connector API

ORI_AUTH_ISSUER

Ori's authorization server (e.g. https://<ref>.supabase.co/auth/v1)

PORT

Local port (default 8787)

How it works

This repository is the MCP server only: a stateless, streamable-HTTP MCP endpoint with JSON replies, built on the official TypeScript SDK. It calls Ori's connector API, which keeps the idea for at most 30 minutes under a one-time token and deletes it when the link is opened. The idea never goes into a URL. Nothing is built and no account is created until you sign up in Ori.

Privacy and terms

Privacy policy · Terms · Support: open an issue in this repository.

This server cannot be deployed

Shareable, User-Owned Context BackendPeople, applications, and agents read, write, and coordinate together in one context lake. Reach it over MCP, a CLI, a Python SDK, or REST.fulcradynamics.comAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Cerebro MCP

    Synvoya
    A
    license
    B
    quality
    D
    maintenance
    Enables non-coders to build software by describing ideas in natural language, orchestrating AI agents and CLI workers via chat without needing a terminal.
    30
    3 npm
    4
    Apache 2.0

  • Slipway

    Sandeepbhuiya29
    A
    license
    A
    quality
    B
    maintenance
    Enables building and shipping real apps directly from Claude or ChatGPT by scaffolding projects, editing code, running live previews with public share URLs, storing data, sending email and push notifications, deploying to hosting, and exporting finished repos.
    22
    MIT

  • Aaly

    candysiva
    A
    license
    Not graded
    quality
    C
    maintenance
    A remote MCP server that lets a coding agent turn a plain-language idea into a real, shareable product — creating the project and data model, serving live API endpoints, deploying custom endpoints and server-side functions, and reading request and function logs. It handles sign-up, OAuth-backed access, and OpenAPI spec generation so people can open, sign into, and use the app the agent builds.
    MIT