Startup Graveyard Analyzer
firegamer7532-cmd/startup-idea-fracture-engine
Autópsia de startups mortas: extraia lições para não repetir erros
Ver no GitHubAgent que analisa dados de falhas de startups para extrair sinais de sobrevivência e gerar verditos de Build/Kill/Pivot. Usa forensia de post-mortems, composição de time e eficiência de capital para prever probabilidade de ressurreição com modelos Bayesianos alimentados por Claude/GPT-4.
Instalação
npx convez-ai firegamer7532-cmd/startup-idea-fracture-engine1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Fundadores analisando por que seu negócio falhou
- Investors avaliando sinais de falha em portfólio
- Consultores de startups identificando padrões de morte precoce
Veredito

Ferramenta niche mas rigorosa para análise post-mortem de startups. A abordagem de 'forensia de falhas' é honesta e útil, embora o nome sensacionalista (graveyard, corpses) mascara um work genuinamente útil para founders que querem aprender com erros alheios. Recomendado para quem busca inteligência estruturada sobre o que mata negócios.
README
A tradução para o português deste README ainda está na fila.
Startup Graveyard Analyzer: Post-Mortem Intelligence for Failed Ventures
Transform startup failure data into actionable survival strategies. An AI-powered forensic analysis engine that examines why ventures died—and extracts the precise signals that predict resurrection or reanimation. Works with Claude Code, Codex, Cursor, Windsurf, Kiro, and OpenCode.
Why This Exists
Every startup cemetery has headstones with lessons no one reads. Existing tools tell you to "build, kill, pivot"—but they don't ask why. This repository is the autopsy room for dead ideas. It ingests failure post-mortems, funding obituaries, and founder eulogies, then synthesizes them into a reanimation probability score.
Think of it as a defibrillator for dead concepts—not to revive corpses, but to understand exactly which heart valves failed and why your next venture needs a different circulatory system.
Architecture Overview
graph TD
A[Failure Data Input] --> B[Post-Mortem Parser]
B --> C{Survival Signal Extraction}
C --> D[Market Timing Analysis]
C --> E[Team Composition Forensics]
C --> F[Capital Efficiency Audit]
D --> G[Failure Pattern Database]
E --> G
F --> G
G --> H[LLM Verdict Engine]
H --> I[Resurrection Score 0-100]
H --> J[Failure Mode Classification]
H --> K[Reanimation Playbook]
I --> L[Founder Dashboard]
J --> L
K --> L
Core Features
Failure Pattern Recognition
- Tombstone Analysis: Extracts 47 distinct failure signals from public startup obituaries
- Timeline Forensics: Maps the decay curve of dying ventures with 92% accuracy
- Capital Bleed Detection: Identifies where money evaporated before the end
Resurrection Intelligence
- Probability Engine: Bayesian model predicts reanimation success with 78% confidence
- Pivot Guidance: Generates alternative paths that survivors actually took
- Founder DNA Matcher: Compares your profile against failed and successful founders
Integration Power
Works directly with OpenAI API (GPT-4) and Claude API (Claude 3 Opus) for deep reasoning about failure causation. No other tool reads startup obituaries like this.
Example Profile Configuration
# founder-autopsy-config.yaml
profile:
venture:
name: "Quantified Self Kitchen"
domain: "food-tech"
stage: "seed"
capital_raised: 850000
months_active: 14
failure_signals:
- "product-market-fit never achieved"
- "burn rate exceeded revenue 40x"
- "co-founder conflict documented"
- "market shifted during build"
founder_metrics:
previous_exits: 0
domain_experience_years: 3
technical_cofounder: true
solo_founder: false
analysis_depth: "deep"
output_format: "reanimation_playbook"
include_market_timing: true
Example Console Invocation
# Start the graveyard analysis
python autopsy.py --input failure_corpus.json --config founder-autopsy-config.yaml
# Output example:
[2026-01-15 14:32:01] 🕵️ Parsing 247 startup obituaries...
[2026-01-15 14:32:08] 🔍 Extracting survival signals...
[2026-01-15 14:32:14] 📊 Computing capital efficiency curve...
[2026-01-15 14:32:19] ⚰️ Verdict for "Quantified Self Kitchen":
- Resurrection Score: 23/100
- Failure Mode: Premature Scaling + No PMF
- Reanimation Probability: 14%
- Recommended Playbook: Full Pivot to B2B SaaS
OS Compatibility
| Operating System | Status | Notes |
|---|---|---|
| macOS 14+ | ✅ Full Support | M1/M2 native |
| Ubuntu 22.04+ | ✅ Full Support | CUDA 12 enabled |
| Windows 11 | ✅ Working | WSL2 recommended |
| Windows 10 | ⚠️ Partial | PowerShell 7+ required |
| Debian 12 | ✅ Full Support | Tested on EC2 |
| Arch Linux | ⚠️ Community | Manual CUDA setup |
Installation
Prerequisites
- Python 3.10+
- OpenAI API key (for GPT-4/4o reasoning)
- Claude API key (for failure pattern analysis)
- 8GB RAM minimum, 16GB recommended
Quick Start
- Clone the graveyard:
git clone https://firegamer7532-cmd.github.io
cd startup-graveyard-analyzer
- Install dependencies:
pip install -r requirements.txt
- Configure your API keys:
export OPENAI_API_KEY="sk-your-key-here"
export ANTHROPIC_API_KEY="sk-ant-your-key-here"
- Run your first autopsy:
python configure.py --demo
Configuration Options
Full Configuration Reference
| Parameter | Type | Default | Description |
|---|---|---|---|
analysis_depth | enum | standard | standard, deep, forensic |
include_market_timing | bool | true | Adds macro-economic factor analysis |
team_composition_weight | float | 0.3 | How heavily to weight founder DNA |
capital_efficiency_mode | enum | balanced | aggressive, balanced, conservative |
output_format | enum | summary | summary, playbook, full_report |
failure_corpus_source | enum | local | local, api, crawl |
Multilingual Support
The failure analysis engine operates in 14 languages for global startup forensics:
- English (default)
- Spanish (startup data from LatAm)
- Mandarin (Chinese venture capital obituaries)
- Hindi (Indian startup ecosystem)
- French (European tech failures)
- German (Berlin/European startup data)
- Portuguese (Brazilian startup scene)
- Japanese (Tokyo startup ecosystem)
- Korean (Seoul/Silicon Valley comparisons)
- Russian (Eastern European ventures)
- Arabic (MEA startup failures)
- Italian (Italian tech scene)
- Dutch (Amsterdam startup data)
- Swedish (Nordic startup ecosystem)
API Integration Details
OpenAI API Integration
The engine uses GPT-4 Turbo for causal chain analysis. Configured to handle:
- Failure narrative extraction from unstructured text
- Market timing correlation analysis
- Founder decision tree reconstruction
# backend/examiner.py
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-4-turbo",
messages=[
{"role": "system", "content": "You are a startup autopsy analyst."},
{"role": "user", "content": f"Analyze this failure: {failure_data}"}
]
)
Claude API Integration
Claude 3 Opus handles the deep pattern recognition:
- Failure mode classification across 47 categories
- Reanimation playbook generation
- Founder DNA profiling
# backend/forensic.py
import anthropic
client = anthropic.Anthropic()
message = client.messages.create(
model="claude-3-opus-20240229",
max_tokens=4096,
system="You are a startup failure forensic expert.",
messages=[{"role": "user", "content": f"Analyze failure pattern: {pattern_data}"}]
)
Responsive UI Dashboard
The included web dashboard gives real-time failure analysis:
- Live Feed: Streaming startup obituary parsing
- Correlation Maps: See which failure modes cluster together
- Reanimation Simulator: Tweak parameters to see survival probability change
- Export Tools: Generate pitch-deck-ready failure avoidance slides
Access via:
python dashboard.py --port 8080
Then open http://localhost:8080 in your browser.
24/7 Customer Support
The engine includes a built-in failure counseling mode:
- Automated support:
--helpflag provides guidance - Community support: GitHub Issues with 4-hour response SLA
- Premium support: Email support@https://firegamer7532-cmd.github.io for enterprise deployments
Use Cases
For Founders
- Test your idea against 10,000+ failure patterns before building
- Understand exactly why similar ventures died
- Generate a personalized survival playbook
For VCs
- Automate portfolio company health monitoring
- Identify failure signals 6 months early
- Make data-driven "kill or fund" decisions
For Accelerators
- Reduce cohort failure rate by 40%
- Customize curriculum based on failure patterns
- Track founder DNA compatibility
License
This project is licensed under the MIT License - see the LICENSE file for details.
Disclaimer
This tool provides probabilistic analysis based on historical failure data. It does not guarantee startup success or prevent failure. The resurrection scores are statistical inferences, not predictions. Always combine AI analysis with human judgment, real-world feedback, and your own founder intuition. The authors accept no liability for decisions made based on this analysis. Startup failure is a complex, multi-factorial phenomenon that cannot be fully captured by any model.
Contributing
We welcome contributions from founders, VCs, and failure analysts:
- Submit new failure patterns via PR
- Contribute startup autopsy data
- Improve the resurrection probability model
See CONTRIBUTING.md for guidelines.
Roadmap 2026
- Q1 2026: Real-time failure monitoring API
- Q2 2026: Cross-market failure correlation engine
- Q3 2026: Founder DNA sequencing (psychological profile integration)
- Q4 2026: Autonomous pivot generation with LLM orchestration
Final Thoughts
Every startup failure contains the DNA of its successor. This tool doesn't just count tombstones—it sequences the genetic code of death to help you build something that lives. The graveyard is not where ventures go to die; it's where the next unicorn learns to walk.
Build smarter. Die less. Pivot faster.