TestimonialInsight - Sales Intelligence from Customer Video Reviews (Sentiment + Objection Tracking)

Upload customer video testimonials, get automatic sentiment scoring, competitor mentions, objection patterns, and ready-to-share highlight clips. Target: SaaS founders who have video testimonials but no way to extract actionable intelligence.

𝕏 Post Reddit HN

Difficulty

beginner

What is it?

SaaS founders collect video testimonials for marketing but use them dumbly — they watch manually, pick highlights by gut feel, and miss the sales intelligence buried inside. TestimonialInsight solves this by uploading a video testimonial, transcribing it, running sentiment analysis, flagging competitor mentions, extracting objection patterns ('price too high', 'too complex', 'implementation time'), and auto-generating highlight clips with subtitles. Users get a dashboard showing sentiment trends across all testimonials, a word cloud of most-mentioned features, and a list of objections they need to address in sales collateral. The core mechanic is video transcription (Whisper) + Claude sentiment/extraction, combined with FFmpeg for clip generation. Why buildable right now: Whisper API is stable, Claude can extract complex patterns from transcripts in a single prompt, FFmpeg works reliably in Lambda/Supabase Functions, and no training is needed.

Why now?

Video testimonials are becoming the standard sales tool (2026 trend), Whisper API just became cheap enough for indie scale, Claude's pattern extraction is now reliable enough for customer insights, and SaaS founders are increasingly data-driven about marketing.

▸Video upload and automatic transcription (Whisper)
▸Sentiment scoring (positive/neutral/negative per speaker turn)
▸Objection extraction (automated flagging of 'too expensive', 'complex', 'implementation time')
▸Competitor mention detection
▸Auto-generated highlight clips with subtitles
▸Dashboard with sentiment trends and feature mentions word cloud
▸Slack alerts for new competitor mentions

Target Audience

SaaS founders and marketing teams collecting 5+ video testimonials (US market: 50k+ B2B SaaS companies, 60% use video testimonials). Focus on post-launch startups with $1k-$10k MRR.

Example Use Case

Sarah, a founder of a project management SaaS, uploads 8 customer testimonials. TestimonialInsight transcribes them, detects that 5 customers mention 'ease of use' and 3 mention competing product names, flags that 2 customers initially worried about pricing but got over it, and auto-generates a 30-second highlight clip of the most enthusiastic customer praising the product's simplicity. Sarah uses this clip in her landing page hero video and adds a FAQ addressing the price objection.

User Stories

▸As a SaaS founder, I want to automatically extract sentiment and objections from customer testimonials, so that I can improve my pitch and address common concerns in sales collateral.
▸As a marketing manager, I want to find and download highlight clips from testimonials without manual editing, so that I can update landing pages faster.
▸As a sales leader, I want to see competitor mentions across all testimonials, so that I can understand what customers are comparing us to and adjust our positioning.

Acceptance Criteria

Video Upload: done when user can upload a .mp4 up to 500MB. Transcription: done when Whisper transcription completes and is displayed within 3 minutes of upload. Sentiment: done when dashboard shows positive/neutral/negative score with confidence. Objections: done when extracted objections are listed with timestamps so user can find them in video. Highlight Clip: done when FFmpeg generates a 30-second clip with correct subtitles and is downloadable.

Is it worth building?

$29/month × 150 paying users (freemium-to-paid at 18%) = $4,350 MRR by month 3. $79/month × 80 users at $29 tier = additional $6,320 MRR by month 5.

Unit Economics

CAC: $8 via Twitter/X outreach (1 hour reaching out yields 3-4 signups, of which 1-2 convert to paid). LTV: $87 (3 months at $29/month, assuming 50% churn after month 2). Payback: 2-3 weeks. Gross margin: 85% (Whisper and Claude are the variable costs).

Business Model

SaaS: free for 1 video/month, then $29/month for 10 videos/month, $79/month for unlimited.

Monetization Path

Free tier: 1 analysis/month. Pro ($29): 10 analyses/month, highlight clip generation, competitor mention alerts. Enterprise ($79): unlimited, team accounts, Slack integration.

Revenue Timeline

First dollar: week 3 via beta upgrade. $1k MRR: month 2. $5k MRR: month 5. $10k MRR: month 10.

Estimated Monthly Cost

Whisper API: $40 (assuming 50 videos/month at $0.008/min, avg 10 min), Claude API: $35 (extraction prompts), FFmpeg processing: $25 (Supabase Functions), Vercel: $20, Supabase: $25, Stripe fees: ~$15. Total: ~$160/month at launch.

Profit Potential

Full-time viable at $4k–$12k MRR.

Scalability

Medium-high — can expand to social media clip generation, testimonial library marketplace, and integration with CRM for lead scoring based on objection patterns.

Success Metrics

Week 1: 300 signups. Week 2: 50 video uploads. Month 1: 40+ paid users. Clip generation success rate: 90%+.

Launch & Validation Plan

Interview 20 SaaS founders currently using video testimonials. Validate: how many testimonials they collect, how they currently extract value, willingness to pay. Build landing page. Recruit 10 beta testers offering free pro tier access for 3 months. Track video upload rate and sentiment accuracy feedback.

Customer Acquisition Strategy

First customer: DM 15 SaaS founders on Twitter/X who have posted customer testimonials, offering free analysis of their existing testimonials in exchange for testimonial about the product. Broader: ProductHunt launch, HackerNews post about video marketing, Twitter/X thread on 'selling yourself via testimonials', Indie Hackers cross-post, LinkedIn content targeting marketing managers.

What's the competition?

Competition Level

Low

What's the roadmap?

Feature Roadmap

V1 (launch): Video upload, Whisper transcription, Claude sentiment extraction, objection flagging, basic dashboard. V2 (month 2-3): Highlight clip generation, competitor mention alerts, sentiment trend charts, Slack integration. V3 (month 4+): Multi-video comparison, social media clip auto-posting, testimonial library marketplace, CRM integration for lead scoring.

Milestone Plan

Phase 1 (Week 1-2): Build landing page, set up Supabase schema, create video upload form, integrate Whisper API, test transcription accuracy. Done when beta user uploads video and transcription appears correctly. Phase 2 (Week 3-4): Add Claude sentiment/objection extraction, build dashboard, integrate FFmpeg for clips, add Stripe checkout. Done when payment flow works and user can download a highlight clip. Phase 3 (Month 2): Optimize FFmpeg performance, add Slack alerts, launch ProductHunt, monitor churn and iterate on objection patterns.

How do you build it?

Tech Stack

Next.js, Claude API, Whisper API, FFmpeg, Supabase Functions, Stripe — build with Cursor for backend video processing, Lovable for dashboard UI, v0 for testimonial cards.

Suggested Frameworks

Time to Ship

14 days

Required Skills

Whisper API integration, Claude API for NLP, FFmpeg command-line scripting, Supabase Functions, React dashboard building.

Resources

Whisper API docs, Claude API docs, FFmpeg documentation, Supabase Functions guide, ProductHunt case studies on video marketing.

MVP Scope

1. Landing page with video upload form (Lovable). 2. Backend endpoint accepting video files and storing in Supabase Storage. 3. Supabase Function triggering Whisper API transcription. 4. Claude API prompt extracting sentiment, objections, competitors from transcript. 5. Real-time dashboard displaying sentiment score and keyword cloud (v0). 6. FFmpeg CLI script generating 30-second highlight clip. 7. Freemium logic and Stripe checkout. 8. Email notification when processing completes.

Core User Journey

Sign up -> upload first testimonial video -> receive sentiment score and key objections in under 2 minutes -> download highlight clip -> upgrade to paid.

Architecture Pattern

Video upload -> Supabase Storage -> Supabase Function triggers -> Whisper API transcription -> Claude API sentiment/extraction -> results stored in Postgres -> dashboard queries Postgres -> FFmpeg generates clip on demand.

Data Model

User has many Testimonials. Testimonial has one Transcription, one SentimentScore, many ExtractedObjections, many HighlightClips. ExtractedObjection belongs to one Testimonial.

Integration Points

Whisper API for transcription, Claude API for NLP extraction, FFmpeg for clip generation, Supabase Storage for video storage, Stripe for payments, Resend for email notifications, Slack API for alerts (future).

V1 Scope Boundaries

V1 excludes: multi-video comparison, social media auto-posting, testimonial marketplace, team collaboration, mobile app, white-label.

Success Definition

A paying SaaS founder uploads a customer testimonial, gets sentiment and objection analysis within 2 minutes, downloads a highlight clip, and reports back that the clip increased their conversion rate or that the objection patterns helped them improve their pitch.

Challenges

Video processing can be slow and expensive at scale. Ensuring transcription accuracy across accents and audio quality. Getting the objection extraction patterns right without false positives.

Avoid These Pitfalls

Do not rely on FFmpeg for real-time clip generation on the frontend — use background jobs via Supabase Functions. Do not over-train on objection patterns without real customer validation — start with rule-based extraction. Do not charge by video file size — it creates perverse incentives for users to upload low-quality videos.

Security Requirements

Auth: Supabase Auth with Google OAuth. RLS: testimonials, transcriptions, and results tables have RLS limiting access to owner. Rate limiting: 5 video uploads/min per IP. Input validation: file size max 500MB, .mp4 only, scan for malware. GDPR: data deletion endpoint removes all videos and transcriptions, document 30-day retention for processing logs.

Infrastructure Plan

Hosting: Vercel (frontend + API routes). Database: Supabase (Postgres). Storage: Supabase Storage for videos. Processing: Supabase Functions for async Whisper/Claude calls. CI/CD: GitHub Actions. Environments: dev (local), staging (Vercel preview), prod (Vercel main). Monitoring: Sentry for errors, Vercel Analytics. Cost breakdown: $160/month includes Whisper, Claude, and Supabase Function invocations.

Performance Targets

Expected load at launch: 100 DAU, 200 video uploads/day. Transcription latency: 10-15 minutes (async, acceptable). Dashboard load time: under 2 seconds. Clip generation time: 30 seconds to 2 minutes. Caching: Postgres query caching for sentiment trends, no aggressive caching for real-time data.

Go-Live Checklist

☐Security audit of video upload and storage (check for XXE, path traversal)
☐Whisper transcription accuracy tested on 10+ diverse accents/audio
☐Claude objection extraction validated against manual labels (target: 85%+ accuracy)
☐Payment flow tested end-to-end (free -> pro)
☐Error tracking (Sentry) live and alerting on API failures
☐Monitoring dashboard showing upload success rate and avg processing time
☐Custom domain testimonial-insight.com set up with SSL
☐Privacy policy and terms published (clarify video retention and processing)
☐5+ beta users signed off with testimonials
☐Rollback plan documented (revert to manual processing if Whisper/Claude fails)
☐Launch post drafted for ProductHunt, HackerNews, Twitter/X, Indie Hackers.

How to build it, step by step

1. npx create-next-app testimonial-insight --typescript. 2. npm install openai @anthropic-ai/sdk @supabase/supabase-js stripe fluent-ffmpeg. 3. Set up Supabase project, create testimonials, transcriptions, sentiment_scores, objections, highlight_clips tables. 4. Build video upload form in Lovable with drag-and-drop. 5. Create /api/upload endpoint storing video in Supabase Storage. 6. Write Supabase Function triggering Whisper API transcription asynchronously. 7. Create Claude prompt extracting sentiment, objections, competitors from transcript. 8. Build dashboard in v0 showing sentiment trends and word cloud. 9. Add FFmpeg script for clip generation (triggered via Supabase Function). 10. Integrate Stripe and Supabase Auth, deploy to Vercel.

Generated

March 28, 2026

Model

claude-haiku-4-5-20251001

← Back to All Ideas