Setup & Operator Guide
Import your workflow, wire two API keys, and ship your first AI-built digital product in under 15 minutes.
Astra
Phase 1 · Strategist
→
Vera
Phase 2 · Builder
→
Orion
Phase 3 · Launcher
Import your workflow, wire two API keys, and ship your first AI-built digital product in under 15 minutes.
Estimated setup time: 10–15 minutes from scratch. No coding required. You'll need accounts for n8n, OpenAI, and ConvertAPI — all have free tiers to get started.
| Requirement | Free? | Where to Get It |
|---|---|---|
| n8n account | Yes — free cloud trial or self-hosted | n8n.io → Sign Up |
| OpenAI API key | Paid — ~$0.10–$0.30 per full run | platform.openai.com/api-keys |
| ConvertAPI secret ⚠️ Required | Free tier — 250 credits/mo (~125 runs) | convertapi.com → Sign Up → Profile → API Secret |
| AI Agent Kit workflow JSON | ✅ Included in your purchase | Your downloads folder |
⚠️ Both API keys are required. OpenAI powers all content + image generation. ConvertAPI converts your ebook and value enhancer into .docx files — without it, the workflow will fail at the final packaging step and you won't get your DOCX output. Set up both before running your first product.
Cost per run: A full Astra → Vera → Orion product run uses approximately 20,000–35,000 tokens. At GPT-4o pricing that's roughly $0.15–$0.30 per product. ConvertAPI's free tier (250 credits/month) covers ~125 DOCX conversions — more than enough to start. At $19 retail, your first sale makes back ~63–125 runs.
Go to your n8n instance (cloud.n8n.io or your self-hosted URL). Click Workflows in the left sidebar.
Click + New Workflow in the top right.
Click the three-dot menu ⋮ in the top-right corner of the canvas → Import from File → select ai-agent-kit-workflow.json from your downloads folder.
Hit Ctrl+S (or ⌘+S on Mac). You should see all 3 agent nodes appear on the canvas connected in sequence.
What you'll see: An orange Form Trigger node on the far left, then three agent clusters — Astra → Vera → Orion — connected in sequence flowing left to right. Any nodes with missing credentials will appear red — you'll fix those in Section 3.
After importing, n8n will highlight nodes with missing credentials in red. You need to wire up two credentials before the workflow can run: OpenAI and ConvertAPI.
Do not skip ConvertAPI. The final two nodes in the workflow — Convert eBook to DOCX and Convert Value to DOCX — use ConvertAPI to generate your .docx files. Without it the run completes but produces no output files. This is the #1 setup mistake.
Used by: all 8 AI nodes (Brainstorming Pains, Generate Solutions, $100M Offer, Create Offer Title, Build Outline, Create $100M Product, Generate Image Cover, Value Enhancer).
Go to platform.openai.com/api-keys → Create new secret key → copy it.
Double-click any red OpenAI node → find the Credential field → click Create new credential → paste your key → Save. Name it something like OpenAI – My Key so it's easy to find.
Click each remaining OpenAI node → Credential dropdown → select the credential you just saved. You only create it once — all other nodes just select it from the list.
Model recommendation: Pre-configured for gpt-4o. You can switch Astra and Orion nodes to gpt-4o-mini to cut costs. Keep Vera (Build Outline + Create $100M Product) on gpt-4o for best manuscript quality.
Used by: Convert eBook to DOCX and Convert Value to DOCX — the two final nodes that package your product.
Go to convertapi.com → Sign up free → you get 250 credits/month at no cost. Each DOCX conversion uses 2 credits.
In the ConvertAPI dashboard → Profile → API secret → copy the string (starts with a long alphanumeric).
Double-click the Convert eBook to DOCX node → scroll to the Parameters section → find the field named Secret (or ApiSecret) → paste your ConvertAPI secret there → Save. Repeat the exact same steps for the Convert Value to DOCX node.
Both Convert nodes should show no warnings. Do a test run — if you see .docx files in the final output node, you're fully wired. If not, re-open the node and confirm the secret field isn't blank.
Free tier is plenty to start: 250 credits/month = ~125 DOCX product runs. That's 125 products per month for free. Upgrade to paid only when you're shipping at volume.
Click the Form Trigger node on the far left. Click Test step to open the live form URL.
Enter your topic/niche (e.g. "Intermittent fasting for busy moms") and target audience. Leave other fields as default for your first run.
Go back to the n8n canvas. Click Execute Workflow. Watch each agent node light up green as it completes. Full run: 3–8 minutes.
Once all nodes turn green, click the last node in the chain (Orion's final output node) → Output tab. You'll find your completed manuscript, cover brief, sales copy, platform listing text, and DOCX download links all in one place.
First-run tip: Use a niche from your Phase 1 Astra Pack (included in your download) — they're pre-validated for demand and profitability. Copy one, paste it in, go.
What she does: Takes your topic, researches real customer pain points, validates demand, frames the offer, and outputs a winning product title + product brief.
What she does: Takes Astra's product brief, writes the full ebook outline, drafts the complete manuscript (3,000–5,000 words), and produces a cover design brief.
What he does: Takes Vera's asset, creates 3 bonuses, writes the full sales page, generates platform listing copy (Gumroad/Etsy/Stan), and outputs the DOCX packaging manifest.
Make sure you're selecting the .json file directly. If n8n shows a parse error, open the JSON in a text editor and confirm it starts with {"nodes":[.
Go to Settings → Credentials in n8n and re-add your OpenAI key. Return to the node and re-select it from the dropdown. If the dropdown is empty, the credential wasn't saved — re-enter it and hit Save again.
The most common cause: the Secret field wasn't saved inside the node. Double-click each Convert node, confirm the Secret field contains your ConvertAPI secret (not blank), and re-run. Also verify your ConvertAPI account has credits remaining (250/mo free).
In n8n cloud, default execution timeout is 60s. Go to Settings → Workflow Settings → set Execution timeout to 300s. For self-hosted n8n, add EXECUTIONS_TIMEOUT=300 to your environment variables and restart.
The #1 cause is a vague input. Revisit your niche brief — add a specific audience, a specific pain, and a specific outcome. Specificity = quality.
Add to Vera's system prompt: "Do not truncate. Write each section fully. Target 500+ words per section."
Still stuck? Email support@aiagentkit.net with your n8n version, which node failed, and a screenshot of the error. We respond within 24h.
Duplicate the workflow in n8n (right-click → Duplicate). Run up to 3 in parallel — each in its own browser tab. Build a product library fast.
After Orion outputs the OTO, run that upsell concept back through Astra as a new product brief. You'll get a complementary product with zero extra research.
Edit any agent's system prompt by clicking the AI Agent node → System Message. Add: "Write in the voice of [your name] — [3 adjectives that describe your style]."
Ask Orion (via a follow-up prompt in the same node): "From this product's outline, give me 10 social media hooks and 5 email subject lines."
Every product the agents build is yours. You can add PLR (Private Label Rights) or resell licenses — Orion's listing copy section has a template for this. Standard pricing: PLR license at 3–5x the base price.