← all drafts

Case Study Candidates — App Dev Offering

Audit of real projects built across GCV, heybrandyn, Mosaic Atlanta, and the Hetzner infrastructure.

Three strongest candidates for the app-dev landing page, with flags on what can go public without client permission.


Candidate 1: REMOVED — Canvass / "community organizing operation"

The original Candidate 1 in this audit was Canvass, framed as a tool built for Mosaic Atlanta. Brandyn flagged this on 2026-05-22 as a hallucinated client relationship: the tool exists but no organization actually uses it in production. Even the anonymized "community organizing operation in Atlanta" framing implies a client deployment that didn't happen. This case study has been removed from all public-facing materials. Do not reintroduce.

If Canvass becomes relevant again later, it should be framed as: "A custom call-analysis tool I built as a side project," with no implied client.


Candidate 2: GCV Content Engine — Sales Funnel + Portfolio Platform

Client: Ghost Coast Video (Brandyn's own business — no permission required)

What was built:

Full automated sales and content system for a video production company:

Problem it solved:

Manual landing page creation, no CRM automation, no content pipeline — just a WordPress site updated by hand. The build turned a "show and tell" website into a lead generation system that qualifies prospects, sends nurture emails, and books calls without manual intervention.

What can be claimed publicly:

Everything. This is Brandyn's own business, built in public. Real numbers: 12 landing pages, 4 case studies, 72 pSEO pages, 77 portfolio assets catalogued, full n8n sales funnel with FluentCRM/FluentBooking.

Flag: None — this is Brandyn's own business. Full detail available.


Candidate 3: heybrandyn Content Automation Engine

Client: heybrandyn (Brandyn's own brand — no permission required)

What was built:

Voice-to-content pipeline with four n8n workflows:

1. Prompts Engine — 636 prompts across 53 weeks, delivered to Telegram with AI-generated spark lines

2. Scheduler — auto-fires Monday and Saturday

3. Telegram Bot — commands for navigating prompts, threading ideas, triggering drafts

4. VoiceNotes Processor — webhook fires when voice note is done, transcribes, runs against 7-part essay framework, returns coverage report + platform-specific copy pulls

One voice recording session → Telegram bot → 7-part essay outline → ~20 pieces of content across 6 platforms.

Problem it solved:

Content production that required full days of writing and editing for a single post. The pipeline collapses that to one recording session, with all downstream formatting automated.

What can be claimed publicly:

Everything. Brandyn is the client and the builder. Real output: 636 prompts, 53-week calendar, voice-to-content loop running live. Do not disclose the specific LLM or API keys used.

Flag: Do not hint at AI use in the content production step — editorial principles apply. The story is "voice note in, content cascade out" without elaborating on what processes the transcript.


Summary Table

| Candidate | Client | Permission needed | Strongest claim |

|---|---|---|---|

| Canvass | Mosaic Atlanta | Yes — for org name | Call analysis tool, community organizing |

| GCV Content Engine | Self (GCV) | No | Full sales funnel, 72-page site, CRM automation |

| heybrandyn Content Automation | Self (heybrandyn) | No | Voice → 20-piece content cascade |

Recommended for landing page: GCV Content Engine and heybrandyn Automation are safe to use in full detail immediately. Use Canvass as a third case study with "community organizing operation in Atlanta" framing until Mosaic permission is confirmed.