The Vercel template that turns a Creator Intelligence System (CIS) instance into a living studio. One-click deploy. Your palace, your voice, your channels. MIT licensed.
GenCreator Studio is a Next.js 16 + Tailwind template. It is the reference gallery, distributor, and attestation surface for CIS — the public face a creator can deploy while authoring remains in Claude Code, Claude Projects, ChatGPT Projects/GPTs, Cursor, or another agent workspace they already use. The protocol lives in creator-intelligence-system; the Studio is what makes it tangible.
┌────────────────────────────────────────────────────┐
│ GenCreator Studio (this repo) │
│ ┌──────────────────────────────────────────────┐ │
│ │ Gallery · Calendar · Formats · Review │ │
│ │ Distribute · Attest · Learn │ │
│ └──────────────────────────────────────────────┘ │
└────────────────────────────────────────────────────┘
│
uses CIP v0.1 protocol
│
┌────────────────────────────────────────────────────┐
│ Creator Intelligence System (sibling repo) │
│ L1 Substrate · L2 Capture · L3 Strategy │
│ L4 Production · L5 Distribution · L6 Learning │
└────────────────────────────────────────────────────┘
- Creators who want a sovereign content brain instead of renting one from Buffer / Notion / HeyGen.
- AI architects who want a reference Vercel template for agent-orchestrated content workflows.
- Independent operators who want to ship a content business without a team but with team-grade rigor.
- One-click Vercel deploy with sensible defaults
- Studio dashboard — Gallery, Calendar, Formats, Review, Distribute, Attest, Learn
- CIS-typed everywhere — every artifact conforms to the CIP v0.1 protocol
- Subscription-first authoring — produce inside Claude Code, Claude Projects, ChatGPT, Cursor, or another creator-owned agent tool
- Attestation rendered on every published piece (the moat)
- Multi-channel distribution — Bluesky, LinkedIn, Beehiiv (MV1) → 8+ channels (MV2)
- Brand voice config — single source via
@cis/voice, banned-phrase + quarantine audit - MIT — fork freely, sell what you build
# 1. Use as template
gh repo create my-studio --template frankxai/GenCreator-Studio --public
cd my-studio
# 2. Install
pnpm install
# 3. Configure
cp .env.example .env.local
# Edit .env.local only when you add distribution channels or optional automation
# 4. Run
pnpm devThen open http://localhost:3000.
GenCreator Studio runs without a server-side LLM key. Environment variables add distribution channels, palace location, and optional hosted automation. See .env.example.
| Variable | Required | Description |
|---|---|---|
AI_GATEWAY_TOKEN |
optional | Vercel AI Gateway for optional hosted automation or vercel-labs/ai-cli workflows |
ANTHROPIC_API_KEY |
optional | Only for hosted server-side LLM automation. Not needed for subscription-first Claude Code/Project authoring |
BLUESKY_HANDLE |
for distribution | App password from bsky.app — free, no API key |
BLUESKY_APP_PASSWORD |
for distribution | Generate at bsky.app/settings/app-passwords |
LINKEDIN_ACCESS_TOKEN |
for distribution | OAuth 2.0 token (60-day) — see DEPLOY.md |
BEEHIIV_API_KEY |
for distribution | From Beehiiv account settings |
CIS_PALACE_DIR |
optional | Path to your SIS palace (defaults to ./.palace/) |
CIS_VOICE_PROFILE |
optional | Path to a custom voice profile JSON |
See DEPLOY.md for:
- One-click Vercel deploy walkthrough
- Custom domain setup
- LinkedIn OAuth setup
- Bluesky app password setup
- Cost estimation (Vercel + optional API automation)
// lib/voice.ts
import { extendProfile, DEFAULT_VOICE } from '@cis/voice'
export const myVoice = extendProfile(DEFAULT_VOICE, {
id: 'my-brand',
northStar: 'Bold. Specific. Earned.',
attributes: {
rigor: 'every claim has a source',
warmth: 'professional without being cold',
},
bannedPhrases: ['journey', 'transformation'],
quarantinedTerms: [],
})The studio loads myVoice at startup and audits every L4 production output against it.
- The Studio is not the authoring engine by default. It displays, distributes, and verifies artifacts produced in the creator's own agent workspace.
- Voice capture is optional. Not the moat. If you prefer to dictate via Claude Code or ChatGPT and import briefs, you can.
- No SaaS infra. This is a deployable template, not a managed service. You run it.
- No AI provider lock-in. Optional automation can use Vercel AI Gateway, provider SDKs, or terminal tools like
vercel-labs/ai-cli; the default path needs none of them. - No auth out of the box. Add
next-author Clerk if you need multi-user. By default, the Studio is a single-creator surface.
- MV1 (this week): Capture surface, Bluesky distribution, attestation rendering
- MV2 (May): All 7 dashboard surfaces, 8 distribution channels, format library
- MV3 (June): AgentDB learning loop, Farcaster Frames, MCP server bundled
- v1.0 (October): Tauri desktop wrapper, multi-creator briefs, music as content type
Full roadmap: creator-intelligence-system/ROADMAP.md.
This template tracks creator-intelligence-system@latest. PRs welcome for:
- New L5 distribution adapters (channel integrations)
- L4 production adapter implementations (image / video / audio providers)
- UI surfaces for the 7 dashboard areas
- Tauri / Electron wrappers
- i18n
See creator-intelligence-system/CONTRIBUTING.md.
MIT. See LICENSE.
- Creator Intelligence System (CIS) — the substrate this Studio runs on
- Starlight Intelligence System (SIS) — L1 substrate (memory palace)
- Library OS — sibling content substrate for books
v0.1.0-alpha · 2026-05-07
Foundation commit. Working flows land across MV1 (this week). Bookmark and watch for v0.1.0 tag.
Open the lid. Speak. Ship. Verify.