Documentation
Everything you need to get the most out of GhostCre8.
Getting Started
What is GhostCre8?
GhostCre8 is an AI-powered content engine that produces brand-disciplined social, email, and editorial content on a weekly cycle. You teach the engine your brand voice once, submit a short brief each week, and the platform generates, quality-checks, and queues your content for publishing. Ghost-created. Brand-approved.
How does the onboarding process work?
After subscribing, you walk through a 5-step guided setup. Step 1 captures your company basics (name, industry, website, audience). Step 2 is the voice interview — you select tone descriptors, set banned terms, define claims boundaries, and provide example sentences that sound (and don't sound) like your brand. Step 3 defines your content pillars with weights. Step 4 connects your publishing channels. Step 5 activates your workspace. The whole process takes 20-30 minutes if you're thoughtful about it — and the more thoughtful you are, the better your content will be from day one.
How long before I see my first content?
After completing onboarding, submit your first brief. The engine generates a full week of content in 8-10 minutes. You'll receive an email notification when the batch is ready for review.
What if I need help with onboarding?
If you'd prefer a professional setup, Nuestream Strategic Partners offers white-glove installation. We conduct a discovery session, extract your voice spec, configure your workspace, and run calibration batches until the output passes your editorial bar. Email drew@nuestream.com to learn more.
The Weekly Brief
What goes into a weekly brief?
The brief has four fields. Theme is required — it's the central topic for this week's content (e.g., 'Why local businesses are invisible to AI search'). Pillar Focus optionally weights ~40% of your posts toward a specific content pillar. Tone Adjustment is a slider that shifts the voice slightly more casual or more authoritative from your default. Special Instructions is freeform — emphasize a specific statistic, feature a product, take a particular angle.
How detailed should my brief be?
More specific briefs produce better content. 'AI visibility for local businesses' is fine but generic. 'Why most HVAC companies score Grade F on AI visibility — lead with the 73% stat, emphasize that traditional SEO doesn't translate, CTA to the free audit' gives the engine much more to work with. Think of the brief as a creative direction, not a topic line.
How many posts does one brief generate?
That depends on your cadence configuration. The default setup generates 12 posts per week: 3 LinkedIn, 3 Instagram, 5 X/Twitter, and 1 blog post. Each post is tailored to the channel's format rules — LinkedIn gets long-form thought leadership, X gets punchy data-driven takes, Instagram gets visual-first captions, and the blog gets a structured article.
Can I submit more than one brief per week?
Yes. Each brief generates a new batch. If you submit two briefs, you'll get two batches to review. Be mindful that more content means more review time and more API costs against your tier limits.
Reviewing Content
What do the QA status indicators mean?
Every piece of content passes through a multi-stage quality assurance check before you see it. Green (Passed) means all checks cleared — the content is ready for your editorial review. Yellow (Warnings) means soft checks flagged something — maybe the tone drifted slightly, a CTA is missing, or the length is outside the channel's ideal range. These are suggestions, not rejections. Red (Failed) means a hard check caught a violation — a banned term was used, a competitor was named, guarantee language appeared, or an unapproved statistic was cited. Failed content is auto-rejected with a specific revision prompt.
How do I approve or reject content?
In the Review dashboard, each content card has Approve and Reject buttons. Approve marks the content as ready for publishing. Reject lets you add revision notes explaining what to change. You can also click Regenerate on rejected posts — this sends the content back through the pipeline with your feedback, producing a new version that addresses the issues.
Can I edit content directly?
Not in the current version. You can copy the content using the Copy button, edit it in your preferred tool, and post it manually. Direct inline editing is on the roadmap.
What's a good first-pass approval rate?
In the first two weeks, expect 50-60% of content to be approved on first pass. By month two, this should rise to 70-75% as the engine learns from your approval patterns. By month three, target 80%+ first-pass approval. If your rate is consistently below 50%, your voice spec likely needs refinement — check the banned terms, tone descriptors, and example sentences in Settings.
Your Voice Spec
What makes a good voice spec?
The voice spec is the most important configuration in your workspace. Three things matter most: tone descriptors (pick 3-5 that genuinely describe how your brand communicates, not how you wish it communicated), banned terms (be specific — generic bans like 'bad words' don't help, but 'leverage, synergy, game-changing, cutting-edge' tells the engine exactly what to avoid), and example sentences (3-5 real sentences that sound like your brand and 3-5 that don't give the engine a concrete target to aim for).
How do I update my voice spec?
Go to Settings, expand the Brand Voice section. You can view your current tone descriptors, banned terms, register rules, claims boundaries, and example sentences. To make changes, update the relevant fields. Changes take effect on the next content generation cycle — they don't retroactively affect existing content.
What are claims boundaries?
Claims boundaries define what your content is and isn't allowed to say. The three defaults are: no guarantee language (never promise specific outcomes), no named competitors (differentiate through capability, not criticism), and additive framing (never disparage what the audience currently has — position your product as the thing they need in addition to what they have, not a replacement). You can customize these in the voice interview during onboarding or in Settings.
What are approved statistics?
If your content frequently cites data (conversion rates, industry benchmarks, survey results), you can add those to your approved statistics list. The QA layer will then verify that every statistic in your generated content matches an approved data point. This prevents the AI from inventing or misrepresenting numbers — every claim survives a literal-truth challenge.
Content Pillars
What are content pillars?
Content pillars are the 3-6 topic categories your brand covers. Each pillar gets a percentage weight that determines how much of your weekly content is allocated to it. For example, a SaaS company might have pillars like Product Education (30%), Industry Trends (25%), Customer Success (20%), Technical Deep Dives (15%), and Founder Perspective (10%).
How do weights work?
Weights should total 100%. A pillar weighted at 25% will get roughly 25% of your weekly posts. If you submit a brief with a specific pillar focus, that pillar gets an additional ~40% boost for that week, with the remaining posts distributed across other pillars by weight.
Can I change pillars after onboarding?
Yes. Go to Settings, expand Content Pillars. You can rename, reweight, add, or remove pillars at any time. Changes apply to the next brief you submit.
Visual Library
What is the visual library?
The visual library is a persistent collection of brand-approved images that the content engine uses when pairing visuals with your posts. Instead of generating a new image for every post (expensive and inconsistent), the engine pulls from your library first and only generates new images when no suitable match exists.
What are reference images?
Reference images are the anchors for visual generation. When the engine does need to create a new image, it uses your reference images as grounding — ensuring the generated image matches your brand's visual identity (colors, style, aesthetic, composition). The more reference images you provide, the more consistent your generated visuals will be.
How do I add images to the library?
In the Library page, click 'Add asset' and paste a Cloudinary URL. Tag the image with relevant categories (product, lifestyle, hero, seasonal, etc.) and optionally mark it as a reference image. Images uploaded to your Cloudinary account can be added here by URL.
Publishing
How does publishing work?
When you approve content in the Review dashboard, it can be published to your connected social platforms. For Instagram and LinkedIn, GhostCre8 uses Postproxy to handle the posting. For blog content, the post is available for copy-paste to your CMS. Direct blog CMS integrations are on the roadmap.
Can I schedule posts for specific times?
Your posting cadence (which days and times each channel posts) is configured during onboarding and can be adjusted in Settings. When content is approved, it's queued according to this cadence. Custom per-post scheduling is on the roadmap.
What if publishing fails?
If a post fails to publish (API error, authentication issue, rate limit), the content remains in 'approved' status and can be retried. The audit log in your workspace tracks all publish attempts. The content itself is never lost — a publishing failure doesn't affect the content asset.
Billing & Plans
What's included in each tier?
Starter ($299/mo): 5 content assets per week, 2 channels, 50-image visual library. Growth ($499/mo): 10 assets per week, 4 channels, 200-image library, 3 team seats, performance tracking. Scale ($799/mo): 20 assets per week, 6 channels, unlimited library, 10 team seats, A/B variants, API access.
Can I change my plan?
Yes. Go to Settings, expand Billing, and click 'Manage subscription.' This opens the Stripe customer portal where you can upgrade, downgrade, or cancel. Changes take effect at the start of your next billing cycle.
Is there a free trial?
We offer beta access codes for early users. Contact drew@nuestream.com if you'd like to try GhostCre8 before committing to a subscription.
Maximizing Value
How do I get the best results from GhostCre8?
Three things matter most. First, invest in your voice spec — the 30 minutes you spend on the voice interview during onboarding directly determines the quality of every piece of content the engine produces. Second, write specific briefs — 'AI visibility for HVAC companies, lead with the 73% Grade F stat, emphasize that nobody in this vertical is doing this yet' produces dramatically better content than 'AI visibility.' Third, review honestly — approve what's genuinely publishable, reject what isn't with clear feedback. Every approval and rejection teaches the engine what you consider good.
How often should I update my voice spec?
Review your voice spec monthly. After 4 weeks of content, you'll notice patterns — certain phrases keep appearing that you don't like, or the tone drifts in a particular direction. Add new banned terms, adjust tone descriptors, and update your example sentences based on what you've approved and rejected. The voice spec is a living document, not a set-and-forget configuration.
What if all my content sounds the same?
This usually means your brief is too narrow or your pillars aren't differentiated enough. Try varying your brief themes week to week, and make sure each content pillar has a distinct angle. You can also adjust the tone slider on your brief to experiment with slightly different registers — more casual one week, more authoritative the next.
Can I use GhostCre8 for multiple brands?
Multi-brand workspaces under a single account are on the roadmap. For now, each brand needs its own GhostCre8 subscription with its own voice spec and configuration. If you manage multiple brands, contact drew@nuestream.com about agency pricing.
Still have questions?
Reach out to Drew directly for help with your workspace configuration, voice spec optimization, or white-glove installation.
drew@nuestream.com