How I built an AI Instagram audit tool and got first paying users in 3 weeks

Hey IH 👋

I want to share the story of building Groly — an AI-powered Instagram audit tool that analyzes any profile in 60 seconds and generates a detailed growth report. We went from idea to paying users in about 3 weeks, and I wanted to share what worked (and what didn't).

The Problem

I was helping a friend with their Instagram strategy and realized something: there's no quick way to get an honest, data-driven assessment of an IG profile. You either hire a social media consultant ($200+/audit), use expensive enterprise tools (HypeAuditor at $299/mo), or piece together free tools that each give you one metric.

I wanted something that could:

• Analyze any public profile (no login required)

• Score it across multiple dimensions (content, engagement, visuals, strategy)

• Give specific, actionable recommendations

• Cost less than a coffee

The Build (2 weeks)

Tech stack:

• Python FastAPI backend

• Celery + Redis for async audit pipeline

• PostgreSQL

• Bright Data API for Instagram scraping

• Gemini Vision for visual content analysis (analyzes actual photos/videos)

• Claude Sonnet for the "5 AI persona" analysis (strategist, content expert, growth hacker, brand consultant, data analyst)

• GPT for EPUB report generation

• Stripe for payments

• Static landing page (no React, no framework — just HTML/CSS/JS)

The audit pipeline works like this:

1. Scrape profile data + last 12 posts via Bright Data

2. Run Gemini Vision on every image/video (visual quality, consistency, color palette)

3. Send everything to Claude with 5 different "persona" prompts — each one analyzes from a different angle

4. Synthesize into a score (0-100) across 5 dimensions

5. Generate a 50-60 page EPUB report with specific recommendations

Total processing time: ~60 seconds.

Key architectural decision: I used multiple AI models for different strengths instead of one model for everything. Gemini is great at vision, Claude is great at nuanced analysis, GPT handles long-form generation well. The cost per audit is about $0.15-0.20 in API calls.

The Pivot (week 2)

I originally built this as a Telegram bot (paid via Telegram Stars). Got ~114 users and 162 audits, but revenue was tiny — about $15 total from Stars payments.

The problem: Telegram users expect everything to be free. Converting them to pay was brutal.

So I pivoted to web-first: built a landing page at groly.dev with a simple flow:

1. Enter any @username → get a free demo (score + 3 insights)

2. See locked premium content (blurred teasers)

3. Pay $4.99 via Stripe → get full 50-page EPUB report by email

The demo acts as a lead magnet and proof of value. You see enough to know the tool works, and the locked content creates FOMO.

Growth Tactics

With zero budget, here's what I did:

SEO pages (week 3):

• Built comparison pages: "HypeAuditor Alternative", "Phlanx Alternative", "NotJustAnalytics Alternative"

• Created free tools that target high-volume keywords: engagement rate calculator (12K/mo searches), hashtag generator (50K/mo), font generator (30K/mo), fake followers checker (22K/mo)

• These free tools funnel into the main audit CTA

Content:

• 3 blog posts targeting "instagram audit tool", "how to analyze instagram profile", "instagram growth strategy 2026"

• Each blog post has CTAs to try the free audit

Social proof:

• Dynamic counter showing real audit count on the landing page

• Live ticker: "🟢 @username just got their audit" (real data from DB)

Numbers (as of March 24, 2026)

- Users: 114 (mostly from Telegram era)

- Audits completed: 162

- Paid reports: 12 (2 real Stripe, 10 from Telegram Stars)

- Revenue: ~$25 total (Stars + Stripe)

- Free tools pages: 6 live

- Blog posts: 3 live, 7 being deployed

- CAC: Effectively $0 so far (no paid ads that converted)

Yeah, the revenue is small. But here's what I'm optimistic about:

1. The SEO play hasn't had time to compound yet — pages are <3 weeks old

2. The free tools are starting to get impressions (targeting 150K+ monthly search volume combined)

3. Agency pack ($39-$149 for bulk audits) just launched — higher ARPU target

4. The product actually works and people share their scores

What I'd Do Differently

1. Start with web, not Telegram. The bot was a fun MVP but Telegram users have near-zero willingness to pay. Should have gone straight to Stripe from day one.

2. Build the free tools first. The hashtag generator and font generator could have been live on day 1 capturing SEO traffic while I built the core product.

3. Price higher. $4.99 is too cheap for the value delivered. Agencies pay $200+/month for similar tools. I'm experimenting with agency packs now.

4. Email capture earlier. I added the drip email sequence late. Should have been there from the start — the 3-email sequence (day 0: results recap, day 2: actionable tip, day 5: paid CTA) is now my best conversion path.

What's Next

• Agency outreach (SMM agencies are the real target — they audit dozens of profiles)

• Product Hunt launch

• Reddit/community seeding

• Potentially raising the individual audit price

Ask

If you're a creator or manage IG accounts, I'd love your feedback. Try a free audit at groly.dev and tell me what's missing or broken.

And if you have questions about the multi-model AI pipeline, Stripe integration, or SEO strategy — happy to go deep in the comments.

---

*Building in public. Revenue goal: get to ramen profitability by April.*