Brainworm
A bookish xAI workspace for Grok chat, code, image generation, and text-to-speech, built with Next.js.
Install
$ agentstack add mcp-h1ddenpr0cess20-brainworm ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ● Environment & secrets Used
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/mcp-h1ddenpr0cess20-brainworm) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Brainworm
A bookish, earth-toned workspace for xAI's Grok — part research companion, part margin scribbler.
Bring your own xAI API key and dig in.
Brainworm is a Next.js app with three native xAI workspaces wrapped in a compact rail, a browser-local thread library, and a manuscript-like feed:
- 📖 Chat — streamed Grok 4.5 responses with reasoning effort, native web search, citation breadcrumbs, reply regeneration and branching, and xAI text-to-speech.
- 🌿 Code — Normal, Plan, and Always-approve session modes, attached source context, user-configured HTTPS MCP workspaces, structured tool activity, and plan approval before implementation.
- 🎨 Imagine — Grok Imagine generation and editing with fast or quality models, aspect ratios, 1K/2K output, and a local image library.
Screenshots
| Chat · Parchment | Chat · Night soil | | :-------------------------------------------------------------------------------: | :---------------------------------------------------------------------------: | | | |
| Code grove · Night soil | Imagine studio · Parchment | | :----------------------------------------------------------------------------------------------: | :-------------------------------------------------------------------------------------: | | | |
Getting started
Requires Node.js 22 or newer (.nvmrc is included).
npm install
npm run devOpen Settings → Model and enter your xAI API key. The same key powers Responses, TTS, voice discovery, and Imagine, and usage is billed to the xAI account that owns it. XAI_MODEL is an optional server environment variable and defaults to grok-4.5.
Prefer a container? The production image uses Next.js standalone output, runs unprivileged, and exposes /api/health as its health check:
cp .env.example .env.local
docker compose up --buildOn your phone: visit the deployed site (e.g. the Vercel deployment) in Chrome and choose ⋮ → Add to Home screen to install Brainworm as a mobile web app. On iPhone, use Share → Add to Home Screen in Safari or Chrome.
Privacy
Everything personal stays in the browser. Your xAI key, threads, settings, and MCP definitions live in localStorage; generated images live in IndexedDB; voice clips live in the Cache API. Credentials are sent only to Brainworm's same-origin API routes, which forward them to xAI for that request without persisting them, and chat requests use store: false.
Safety and intended use
Brainworm is a creative and technical tool. It is not intended for children and must not be used as a companion, friend, therapist, caregiver, or romantic partner. AI output can be inaccurate, unsafe, or inappropriate; never rely on Brainworm for crisis response, professional advice, or safety-critical decisions.
Read the [AI Output Disclaimer and Conditions of Use](docs/ai-output-disclaimer.md) and [Not a Companion](docs/not-a-companion.md) policy before using or deploying the app. The [documentation index](docs/README.md) summarizes these boundaries.
MCP workspaces
Add up to eight remote HTTPS MCP servers under Settings → Workspaces. Each server has two exact tool allowlists:
- Read-only tools are exposed in Normal and Plan modes.
- Always-approve tools are exposed only after you select Always-approve or approve a plan for implementation.
An empty allowlist exposes no tools. Server labels are normalized and deduplicated, insecure URLs are rejected, and authorization headers stay browser-local between requests.
Development
npm run check # format:check + lint + typecheck + test:coveragenpm run format and npm run lint:fix apply safe automated fixes; GitHub Actions repeats the same quality gate on every pull request.
References
- Darkwords — product structure and responsive interaction reference
- Wordmark — xAI TTS and Grok Imagine behavior reference
- xAI Grok Build — agent modes, plan approval, tool visibility, sessions, and permission workflow reference (Apache-2.0)
- xAI Responses API · Text to Speech · Image Generation · Remote MCP
Created with GPT-5.6 Sol · extra high · fast
Source & license
This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: h1ddenpr0cess20
- Source: h1ddenpr0cess20/brainworm
- License: Apache-2.0
- Homepage: https://brainworm-nu.vercel.app
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.