# Cadence

> From spark to scheduled post: Cadence finds topics, runs agentic AI research, writes in your voice, and publishes automatically to LinkedIn, Blog & WordPress.

Cadence is built for two audiences. **Solo creators and thought leaders** who want a steady LinkedIn presence without becoming their own content manager, and **social media managers and agencies** who want to scale client output without scaling working hours.

The platform is hosted in the EU, GDPR-compliant, and built by GOLLE Digital Solutions GmbH (Germany).

## Languages

- English: https://cadence.golle.io/en
- German (default): https://cadence.golle.io/

## Core capabilities

- AI-powered topic discovery from company profile, industry RSS feeds, and current trends
- Agentic multi-tool research with query planning, source curation, and a reflection loop (the most expensive phase per pipeline item; researchers across multiple tools rather than a single search)
- Multi-channel content generation in a single run: blog articles (Markdown), LinkedIn posts (LinkedIn-accurate UTF-16 character counter), WordPress posts (HTML, multi-site)
- AI hero images at the right aspect ratio per channel, with brand consistency via uploaded reference people and products
- Per-workspace writing-style training from 200--15,000 character samples (tone, formality, perspective, sentence structure, signature phrases, do/don't rules)
- Automatic translation to configured target languages with channel-specific guidelines and SEO-aware phrasing
- Smart scheduling per channel (best weekday, automatic character-budget enforcement, "Fix with AI" shortener for over-budget LinkedIn posts)
- Direct publishing to LinkedIn (personal and company pages, OAuth 2.0), native blog, and WordPress (multi-site, application-password auth, no plugin needed)
- Multi-workspace architecture for agencies: one account, any number of client workspaces, role-based access (Owner / Admin / Editor / Viewer), per-client writing style and brand assets
- Frontier LLMs in max-thinking mode: Cadence is LLM-agnostic; the strongest current frontier models are validated and rolled out to every workspace automatically

## How the pipeline works

1. **Ideation** -- Drop an idea — text, URL, or PDF. Or let the Discovery agent suggest fresh topics.
2. **Analysis** -- AI scores your idea (1–10), suggests channels, tone, and audience — refinable in an interactive chat.
3. **Research** -- Agentic multi-tool research: the agent plans queries, scans the web, curates sources, spots gaps, and re-researches — like a senior researcher would.
4. **Creating** -- Multi-channel generation in a single run: blog article, LinkedIn post, plus AI-generated hero image at the right aspect ratio.
5. **Review** -- Inline editor + AI feedback chat per deliverable. “Fix with AI” shortens over-budget LinkedIn posts in one click — without losing hook, hashtags, or CTA.
6. **Translation** -- Optional automatic translation into all configured languages — with channel guidelines and SEO phrasing per market.
7. **Publishing** -- Direct to LinkedIn, Blog, or WordPress — with smart scheduling that suggests the optimal weekday per channel.

After Publishing, posts can be scheduled for the optimal weekday per channel. The full pipeline (Ideation -> Analysis -> Research -> Creating) can run end-to-end in the background as Auto-Pilot; only the Review step requires a human.

## Channels supported

- **LinkedIn** -- Direct publishing to profiles & company pages. OAuth connection to personal & company pages; Live char counter (LinkedIn-accurate, UTF-16); Automatic “Fix with AI” shortener when over-limit; Smart link strategy (blog link, CTA link, or no link); Hero image in 4:5, smart scheduling.
- **Blog** -- Native blog posts with multi-language. Auto-generated slugs & meta descriptions; Hero image in 3:2; Multi-language variants under one slug; Scheduled publishing & revisions; Full SEO optimisation.
- **WordPress** -- Multi-site publishing. Every WordPress site = its own channel; Auth via application passwords (no plugin needed); Auto featured image, default author/categories; Pretty- and query-permalinks supported; Background cron with retry logic.

## What makes Cadence different

- **Frontier LLMs in max-thinking mode** -- We use only the latest top-frontier models (currently Claude by Anthropic for analysis, research and text, Gemini by Google for images) at the highest reasoning level. Cadence is LLM-agnostic: as soon as a stronger model arrives, we validate and auto-upgrade every workspace. Maximum intelligence — no model management on your side.
- **Agentic multi-tool research** -- Our research agents work in loops: plan queries, scan the web, score sources, spot gaps, re-research. With a variety of tools — not a single Google search, but real investigative research, the way a senior researcher would do it.
- **Brand consistency you don't have to brief** -- Upload sample texts (200–15,000 chars) and Cadence extracts your writing-style profile: tone, sentence structure, signature phrases. Upload people and product images and AI-generated hero images stay on-brand. Multiple clients? One profile per workspace.
- **Platform-perfect down to the detail** -- LinkedIn-accurate character counter (UTF-16, bold and emojis count double), automatic "Fix with AI" shortener, SEO-optimised slugs and meta descriptions, automatic translations with channel-specific guidelines. Posts go live without you knowing platform quirks.

## Pricing (EUR, monthly)

- **Free** (0 €/month) -- To dip your toe in. 0 credits/month; 1 channel; 1 profile; 1 user; Top-up: 20 € / 1,000 credits.
- **Starter** (29 €/month) -- For solo creators. 2,500 credits/month; 5 channels; 3 profiles; 3 users; Email support; Top-up: 15 € / 1,000 credits.
- **Pro** (79 €/month) -- For agencies & power users. 8,000 credits/month; 15 channels; 10 profiles; 10 users; Priority support; Top-up: 10 € / 1,000 credits.
- **Enterprise** (custom) -- For large teams. Unlimited credits*; Unlimited channels; Unlimited profiles; Unlimited users; Dedicated support & SLA; Custom integrations.

Top-up credits: 20 EUR per 1,000 credits on Free, 15 EUR on Starter, 10 EUR on Pro. Median real-world cost per pipeline item is approximately 165 credits (production data); research is the most expensive phase.

## Sign up

https://cadence.golle.io/auth/signup

## Contact

cadence@golle-it.de -- GOLLE Digital Solutions GmbH (https://golle-it.de)

## Machine-readable references

- Sitemap: https://cadence.golle.io/sitemap.xml
- Robots: https://cadence.golle.io/robots.txt
- Full content: https://cadence.golle.io/llms-full.txt

## Frequently asked (English)
### Does Cadence replace human content managers?

No. Cadence handles the heavy lifting — research, drafting, translation, publishing — so content managers can focus on strategy, clients, and creative decisions. Humans stay in the review loop.

### Which AI models does Cadence use?

We exclusively use the latest top-frontier models (currently Claude by Anthropic for analysis, research and text, Gemini by Google for images) at the highest reasoning level. Cadence is LLM-agnostic: as soon as a stronger model is available, we validate it briefly and auto-upgrade every workspace.

### How does the research work?

Cadence runs agentic multi-tool research: the agent plans queries, scans the web with various tools, scores sources, spots gaps, and re-researches in a reflection loop — until enough reliable sources are found. The result is a structured briefing with all sources.

### Can I train my writing style?

Yes. Upload 200 to 15,000 chars of your own samples and Cadence extracts a structured writing-style profile: tone, formality, perspective, sentence structure, vocabulary, signature phrases, do/don't rules. The profile is automatically applied to every generation.

### Is my data GDPR-compliant?

Yes. Cadence is hosted in the EU, GDPR-compliant, and offers per-customer workspace isolation. The data processing agreement under Art. 28 GDPR is concluded automatically with every contract - on every plan, at no extra cost and without a signature. At Enterprise level, individual hosting setups and a signed counterpart are added.

### Which channels are supported?

LinkedIn (personal profiles and company pages), Blog (native posts with multi-language and hero image), and WordPress (multi-site, every site as its own channel). More channels are coming.

### Is there an API?

Yes. Cadence offers a full REST API with OpenAPI documentation — including SSE streams for live updates. Available from the Pro plan.

### What happens when my credits run out?

You can top up credits at any time (1,000 credits, price depends on plan). Active pipeline items pause until enough credits are available. On plan renewal, credits are topped up to the plan level.
## Häufige Fragen (German)
### Ersetzt Cadence menschliche Content-Manager?

Nein. Cadence übernimmt die Fleißarbeit – Recherche, Erstellen, Übersetzen, Publishing – damit Content-Manager sich auf Strategie, Kunden und kreative Entscheidungen konzentrieren können. Der Mensch bleibt im Review-Loop.

### Welche KI-Modelle nutzt Cadence?

Wir setzen ausschließlich auf die jeweils aktuellsten Top-Frontier-Modelle (derzeit Claude von Anthropic für Analyse, Recherche und Text, Gemini von Google für Bilder) im höchsten Reasoning-Modus. Cadence ist LLM-agnostisch: Sobald ein neues, stärkeres Modell verfügbar ist, validieren wir es kurz und upgraden alle Workspaces automatisch.

### Wie funktioniert die Recherche?

Cadence führt agentische Multi-Tool-Recherche durch: Der Research-Agent plant Suchanfragen, durchsucht das Web mit verschiedenen Tools, bewertet die Quellen, erkennt Lücken und recherchiert in einem Reflection-Loop nach – bis genug verlässliche Quellen vorliegen. Das Ergebnis ist ein strukturiertes Briefing mit allen Quellen.

### Kann ich meinen Schreibstil trainieren?

Ja. Lade 200 bis 15.000 Zeichen eigene Sample-Texte hoch und Cadence extrahiert ein strukturiertes Schreibstil-Profil: Tonalität, Formalität, Perspektive, Satzbau, Vokabular, Signature-Phrasen, Do/Don't-Regeln. Das Profil wird automatisch bei jeder Generierung angewendet.

### Sind meine Daten DSGVO-konform geschützt?

Ja. Cadence wird in der EU gehostet, ist DSGVO-konform und bietet Workspace-Isolation pro Kunde. Der Auftragsverarbeitungsvertrag nach Art. 28 DSGVO wird automatisch mit jedem Vertrag geschlossen – in jedem Tarif, ohne Zusatzkosten und ohne Unterschrift. Auf Enterprise-Niveau kommen individuelle Hosting-Setups und eine unterzeichnete Ausfertigung hinzu.

### Welche Channels werden unterstützt?

LinkedIn (Privat-Profile und Company-Pages), Blog (native Posts mit Multi-Sprache und Hero-Image) und WordPress (Multi-Site, jede Site als eigener Channel). Weitere Channels sind in Vorbereitung.

### Gibt es eine API?

Ja. Cadence stellt eine vollständige REST-API mit OpenAPI-Dokumentation zur Verfügung – inklusive SSE-Streams für Live-Updates. Verfügbar ab dem Pro-Plan.

### Was passiert, wenn meine Credits aufgebraucht sind?

Du kannst jederzeit Credits per Top-Up nachladen (1.000 Credits, Preis abhängig vom Plan). Aktive Pipeline-Items pausieren bis ausreichend Credits vorhanden sind. Bei Plan-Renewal wird auf das Plan-Niveau aufgefüllt.
## Hero positioning, English
### For solo creators / thought leaders
**Your content. On autopilot.**
Cadence finds topics, researches deeply, writes in your voice, and publishes automatically — on LinkedIn, your blog, everywhere.
### For social media managers / agencies
**5 clients? 50 clients? Same hours.**
The content pipeline for social media managers: topics, research, copy, images, publishing — per client with their own voice and channels.
## Hero positioning, German
### Für Solo-Creator / Thought Leader
**Dein Content. Auf Autopilot.**
Cadence findet Themen, recherchiert tief, schreibt in deinem Stil und veröffentlicht automatisch – auf LinkedIn, im Blog, überall.
### Für Social-Media-Manager / Agenturen
**5 Kunden? 50 Kunden? Selbe Stunden.**
Die Content-Pipeline für Social-Media-Manager: Themen, Recherche, Texte, Bilder, Publishing – pro Kunde mit eigener Stimme und eigenen Channels.
## Workflow patterns
### Solo workflow (English)
- One profile with your writing style and topics
- Discovery delivers ideas in the morning — pick one, click “Auto-Pilot”
- By afternoon the post is review-ready
- Smart scheduling makes sure you show up at the right time
### Agency workflow (English)
- Multi-workspace: 1 account, any number of client workspaces
- Per client: own profile, own style, own channels, own branding
- Roles: Owner / Admin / Editor / Viewer (clients approve read-only)
- Audit logs & multi-user collaboration
Agency math: At 500 €/client, 10 clients today → 4× scaling with Cadence (10h → 2h per client, same hours). Estimated upside: +15,000 €/month.
## The patchwork problem (why Cadence exists)
Today nearly everyone writes copy with ChatGPT, Claude, or Gemini. The problem isn't the writing — it's everything around it.
Today's typical AI-content workflow:
- Find ideas worth posting in the first place
- Hunt down news and industry insights
- Run deep research in yet another tool
- Shape it into a structure
- Analyse and sharpen the idea
- Have the content written
- Translate for other markets
- Generate images for it
- Edit — sometimes multiple times
- Manually copy-paste into the channel
- Realise: “Damn, the LinkedIn post is 3,200 chars.”
- Post at the right time (or forget)
With Cadence:
- Drop in an idea — or let the AI suggest one
- Cadence handles it: analysis, research, content, images (photo, infographic, quote card …), translation
- You review — we publish, scheduled to the perfect time
Even with AI, you still spend hours per content piece. No autopilot. No flow. That's exactly where Cadence steps in: a single pipeline from spark to scheduled post.