| What it does | AI research tool that watches winning short-form videos and helps Claude and other AI assistants generate creative hooks, scripts and strategies
| Run structured AI debates, pitch reviews, focus groups, audience simulations and idea tournaments with inspectable transcripts and reports
|
|---|
| Stage | Startup
| MVP
|
|---|
| Founded | 2025
| Not given
|
|---|
| Based in | KE
| Not given
|
|---|
| Pricing model | Paid
| Not given
|
|---|
| Pricing | Pay per Run in prepaid Credits at 1,000 credits per US dollar. Research of 50 videos: 1,036 credits (~$1.04). Analysing 20 videos: 460 credits (~$0.46). Tracking 10 creators daily for a month: 1,800 credits (~$1.80). New accounts get 1,000 free credits valid for 90 days. Top-ups from $10 (10,000 cre
| Not given
|
|---|
| Key features | - Works with Claude.ai, ChatGPT, Cursor, VS Code and any MCP-compatible assistant
- Access public data from TikTok, Reels, Shorts, Facebook, Meta Ads and creator metrics
- Score videos by how far they outperform a creator's usual reach
- Pay per Run with no subscription or monthly seats
- Free reading of completed Runs and all research results
- Supports API access via REST endpoints for developers
| - Decision stress-test with opposing agents and convergence recommendation
- Judged debate with neutral arbiter scoring
- Expert panel with key insights without verdict
- Focus group with moderated discussion guide and verified quotes
- Idea tournament with bracket and kill cards
- TribeMind synthetic audience with opinion tracking
- Shark Tank pitch panel with cross-examination
- Interview simulator for testing guides on simulated participants
- MultiFormat workflow chaining multiple formats together
- Support for major AI models including Anthropic, OpenAI, Google Gemini, and others
|
|---|
| What makes it different | - Brings video data to the AI you already use instead of requiring a new app or subscription
- Analyzes actual video content including hook, visuals and call-to-action, not just captions
- Pay only for work that completes; failed runs cost nothing
- Ranks videos by outperformance relative to creator baseline, not absolute views
| - Structured protocol determines turn order, not model inference
- Each participant has separate briefs and private context
- Persistent transcripts preserved turn by turn
- Human steering mid-discussion without restarting
- Multiple dissenting AI perspectives instead of single framing
|
|---|
| Community votes | 0 | 0 |
|---|
| DR (Domain Rating by Ahrefs) | 53 ▲ +1 (up) since last check | 4 ● no change since last check |
|---|
| Trust Flow (Majestic) | 17 | 0 |
|---|
| Citation Flow (Majestic) | 31 | 11 |
|---|
| Referring domains (Majestic) | 130 | 1 |
|---|
ShortsIntel
Delibora