Multi-LLM content strategy dashboard that turns your business context, personas, and weekly brief into a research-grounded social calendar for X, LinkedIn, and Reddit.
Plan the week once. Generate platform-native drafts with hooks, rationale, and optional images. Log performance and feed it back into the next run.
Weekly content calendar with persona and platform filters, generation status, and slot cards by day.
Track posts through draft → approved → posted → logged.
Review hooks, hook variants, body, rationale, and citations for each slot.
Paste a brain dump and generate six platform variants (LinkedIn / X / Reddit × polished / faithful).
Set the week's focus and manage voices with platform cadence.
Configure AI gateway connectivity and per-task model routing.
Auren is a single-user content operating system for founders and small teams who publish under multiple voices. You configure strategy once (ICP, goals, personas, story bank), set a weekly brief, and run a generation pipeline that:
- Researches fresh signals (trends, competitors, news) grounded in your ICP and region
- Analyzes performance from past posts to learn what worked
- Plans the week with a mixed content calendar (research, story, fun, opinion)
- Writes each slot through a multi-step LLM chain: draft → hook critic → polish → rationale
- Surfaces drafts on a calendar or kanban board for review, approval, and export
There is also a Rewrite mode: paste a raw brain dump and get six platform variants (LinkedIn / X / Reddit × polished / faithful), optionally fact-checked.
| Area | What you get |
|---|---|
| Calendar | Week view and kanban by post status (draft → approved → posted → logged) |
| Strategy | Business profile, quarterly goals, personas with voice + cadence, story bank, weekly brief |
| Generation | Research → strategy → write pipeline with per-task model routing |
| Rewrite | Brain-dump → multi-platform social variants with optional fact-check |
| Performance | Import metrics and qualitative notes; patterns feed the next generation |
| Images | Optional image generation for posts via your AI gateway |
| Settings | Per-task LLM routing, content mix ratios, generation concurrency |
Each weekly run moves through staged jobs you can watch in the UI:
Research → Performance digest → Strategy plan → Slot writing (parallel) → Done
Content mix defaults to 50% research, 20% story, 15% fun, 15% opinion. Story and opinion slots can draw from your story bank (first-person anecdotes and hot takes).
Default model routing (all overridable in Settings):
| Task | Default model |
|---|---|
| Research | Perplexity Sonar Pro (web search, last-week recency) |
| Strategy | Claude Sonnet 4.5 |
| Write | Claude Sonnet 4.5 |
| Hook critic | GPT-5.5 |
| Polish | Claude Haiku 4.5 |
| Rationale | Claude Sonnet 4.5 |
| Feedback analysis | Claude Opus 4.7 |
Every post ships with structured rationale (hook strategy, audience, slot reasoning, virality lever) and optional citations back to research signals, personas, goals, or performance data.
- Next.js 16 (App Router) + React 19 + Tailwind CSS v4
- SQLite via Drizzle ORM + libSQL (local file or Turso)
- Bun for dev scripts and DB tooling
- AI API Gateway for multi-provider LLM routing, web search, and image generation
- Bun (recommended) or Node.js 20+
- A running AI API Gateway instance with at least one chat provider configured (OpenRouter recommended for the default routing)
- Optional: Turso account for hosted SQLite in production
git clone https://github.com/muneebhashone/auren-content.git
cd auren-content
bun installcp .env.local.example .env.localEdit .env.local:
# AI API Gateway — owns provider credentials, CLI auth, web search, and model discovery
AI_GATEWAY_BASE_URL=http://127.0.0.1:3000
# Optional image generation defaults
AI_GATEWAY_IMAGE_MODEL=codex/gpt-5.5
AI_GATEWAY_IMAGE_SIZE=1024x1024
AI_GATEWAY_IMAGE_RESPONSE_FORMAT=b64_json
# Database — local file for dev, Turso URL + token for production
TURSO_DATABASE_URL=file:./local.db
TURSO_AUTH_TOKEN=
# Optional HTTP Basic Auth gate (set both to enable)
DASHBOARD_USER=auren
DASHBOARD_PASSWORD=your-secretbun run db:migrate
bun run db:seed # optional: sample business profile + personaRun your AI gateway (see ai-api) on the port configured above, then:
bun run devOpen http://localhost:3000.
The dev script clears stale running generation jobs on startup so interrupted runs do not block new ones.
- Settings — confirm your gateway is reachable; adjust per-task model routing if needed
- Business Profile — ICP, services, brand pillars, voice
- Quarterly Goals — activate one objective for the current quarter
- Personas — add voices with platform cadence (e.g. 3× X, 2× LinkedIn per week)
- Story Bank — seed anecdotes and hot takes for story/opinion slots
- Weekly Brief — set focus + target segment, run research, then generate the week
bun run dev # dev server (clears stuck jobs first)
bun run build # production build
bun run start # production server
bun run lint # ESLint
bun run db:generate # Drizzle migration files
bun run db:migrate # apply migrations
bun run db:seed # seed sample strategy data
bun run db:studio # Drizzle Studioapp/
(dashboard)/ # Calendar, rewrite, performance, strategy, settings
api/ # Generation, rewrite, and image routes
components/ # Calendar, post detail, nav, UI primitives
lib/
db/ # Schema, migrations, client
generation/ # Pipeline, content mix, brief suggester, humanize
llm/ # Gateway client, router, prompts
images/ # Image generation helpers
scripts/ # Dev bootstrap, seed utilities
- Set
TURSO_DATABASE_URLandTURSO_AUTH_TOKENfor a hosted libSQL database - Point
AI_GATEWAY_BASE_URLat your deployed gateway - Set
DASHBOARD_USERandDASHBOARD_PASSWORDbefore exposing the dashboard publicly - Generated images are saved under
public/generated/
MIT