What It Costs To Run QWF

What It Costs To Run QWF

This page exists because we believe donor-partners deserve to see exactly where every dollar goes… not a polished pie chart, the actual line items.

Every figure on this page is generated automatically from the same constants file our backoffice scripts use. When infrastructure changes, this page changes. When we update a constant, an audit runs the next morning and reconciles it against the actual Azure invoice. If they disagree by more than 5%, the system raises an alert before anyone can publish a wrong number.

Last regenerated: 2026-04-26 (Pacific). Constants verified: 2026-04-25.


The Headline

  • Fixed monthly infrastructure: $378.54
  • Variable monthly budget (LLM + scraping): up to $400.00
  • Grand total at full variable spend: up to $778.54/month

That funds 13 active apps, ~50 automation workflows, ~280 Python scripts, and the entire backoffice that keeps QWF running.

Most recent month-to-date Azure spend (live): $175.57 as of 2026-04-26.


Azure (cloud compute, storage, networking)

Two virtual machines + supporting infrastructure. The big one runs the backoffice; the small one runs n8n workflow orchestration.

ResourceMonthly CostWhat it does
claude-dev compute (D8as_v6, 8 vCPU, 32 GB)$170.82Main backoffice VM… runs all Python scripts, hosts the Digital Twin API, manages the Obsidian vault that is QWF’s brain
claude-dev disk (128 GB NVMe Premium SSD)$27.00Storage for the vault, databases, and execution environment
qwu-n8n compute (B2s, 2 vCPU, 4 GB)$30.00n8n workflow orchestration (~50 active workflows for emails, scheduling, monitoring)
qwu-n8n disk (Standard SSD)$4.00n8n workflow + execution data
2x static public IPs$8.00One per VM, $4 each
Pre-resize rollback snapshot$0.72Safety net… deletable after stability confirmed
Azure subtotal$240.54

Supabase (databases + auth + edge functions)

One Supabase organization on the Pro plan. 12 apps each get their own MICRO-tier project; 1 app runs on the Free tier.

ComponentMonthly CostWhat it does
Supabase Pro org base$25.00Single org covering all QWF apps
Included compute credit($10.00)Pro plan credit (offsets one MICRO project)
12 MICRO projects @ $10.00 each$120.00One per app: QWR, HQ, QQT, QRP, QSP, QNT, QKN, Pocket EZ, L4G, QTR, WHL, GreenCal CC
Supabase subtotal$135.00

Other Fixed Infrastructure

ServiceMonthly CostWhat it does
RackNerd VPS (ESP)$3.00Self-hosted email service provider on a $35/year VPS that actually works (annual: $35.49)
Betterstack monitoring$0.00Lifetime AppSumo license… no recurring cost
Email (M365)$0.00Included in existing M365 license

Variable Costs (Budgets, Not Floors)

Two things scale with use: AI model calls (LLM) and external data scraping (Apify). These are budgets we monitor daily; actual spend is usually well below.

BucketMonthly BudgetWhat it pays for
OpenRouter (LLM API)$300.00Claude + other models for content analysis, automation, research
Apify (web scraping)$75.00Lead generation, public data collection (LinkedIn, maps, etc.)
Variable budget total$400.00

If we hit 75% of either budget, the system pings a Discord alert. If we hit 90%, that becomes a critical alert. We have not exceeded these budgets at the monthly level.


The Bottom Line

Fixed: $378.54/month. Maximum if every variable budget hits 100%: $778.54/month.

That covers every app, every automation, every dashboard, every email, every database, every monitor. No hidden line items. No marketing budget hidden in ‘operations’. No executive compensation buried in ‘professional services’.

If you want to see exactly which services map to which programs, see The Tool Shed for the full vendor-by-vendor breakdown of all 48+ tools we use.


How This Number Stays Honest

Three things keep this page from drifting into marketing fiction:

  1. Single source of truth. Every dollar figure on this page comes from cost_constants.py in our public-by-default backoffice. The Digital Twin, internal dashboards, audit reports, and this transparency page all read the same constants.
  2. Daily reconciliation. A scheduled audit pulls our actual Azure spend each morning and compares it to the constants. Drift over 10% raises an alert.
  3. Monthly invoice reconciliation. On the 8th of each month, after Azure’s invoice settles, we pull the finalized total for the prior month and reconcile it against the constants with a tighter 5% threshold.

Most recent monthly reconciliation: 2026-03 (internal report; full backoffice is browseable on GitHub)

Source script: 005 Operations/Execution/generate_cost_transparency.py Constants file: 005 Operations/Execution/cost_constants.py Daily audit: 005 Operations/Execution/audit_system.py (cron 6 AM Pacific) Monthly reconciliation: 005 Operations/Execution/reconcile_azure_invoice.py (cron 9 AM Pacific on the 8th)