Oratruth

## 🏛️ LatAm Truth Oracle

The Bloomberg Terminal for Latin America’s Real Economy.

### 🔴 The Problem

Latin America faces a massive trust crisis in economic data. Official inflation stats are often manipulated, delayed, or blind to the informal markets where 50% of the population shops.

### 🟢 Our Solution

We’ve built a Bittensor subnet that incentivizes miners to verify physical prices on the ground.

- Proof of Physical Truth (PoPT): Miners must provide OCR-verified photos of prices alongside unique cryptographic challenge hashes.

- Commit-Reveal Protocol: Prevents front-running and data plagiarism between miners.

- Gaussian Consensus: Automatically slashes outliers and attackers, rewarding only truthful data.

### ⚙️ Technical Innovation

1. Accuracy (60%): Statistical consensus using standard deviation filters.

2. Proof (30%): EasyOCR integration to validate physical presence.

3. Latency (10%): Rewarding high-speed data delivery.

### 🚀 Roadmap

- Phase 1: Local simulation and architecture validation (Current).

- Phase 2: Testnet deployment with miners in AR, CL, and BR.

- Phase 3: API commercialization for Fintechs and Remittance platforms.

# 🏛️ LatAm Truth Oracle — Pitch Deck

## 10-Page Business Presentation

---

## SLIDE 1: COVER

# LatAm Truth Oracle
### Verifiable Real-World Price Data on Bittensor

*Subnet Ideathon — Round I Submission*
*February 2026*

**Tagline:** *"What if you could verify the price of bread in Buenos Aires without trusting anyone?"*

---

## SLIDE 2: THE PROBLEM

# 🔴 Latin America Has a Data Trust Crisis

### 700 million people can't trust their own price data

- **Argentina**: Government manipulated inflation by 2-3x for 8 years (INDEC scandal)
- **Venezuela**: Official exchange rate diverged 1000x from reality
- **Remittances**: $150B/year sent to LatAm — senders don't know what recipients can buy

### The Legacy Oracle Problem:
- 📅 Government stats are **manipulated**
- 🌐 Web scrapers are **blind to physical reality**
- 🏦 Crypto oracles only see **digital tokens**

> **There is NO real-time, physically-verified consumer price data for Latin America.**

---

## SLIDE 3: THE SOLUTION

# 🟢 Decentralized Price Verification Network

### LatAm Truth Oracle: A Bittensor Subnet

Miners **physically observe** prices at local markets and submit data through a **cryptographic commit-reveal protocol**.

```
🔒 COMMIT   →  Miner submits SHA256(price) blindly
⏳ 60s WAIT  →  Anti-gaming timeout (no copying)
🔓 REVEAL   →  Miner reveals data + photo proof
🏆 SCORE    →  Gaussian consensus + OCR verification
```

**Result:** Trustworthy, real-time price index that no single entity can manipulate.

---

## SLIDE 4: HOW IT WORKS

# ⚙️ The Commit-Reveal Protocol

### Three-layer verification prevents all known attack vectors:

| Layer | Mechanism | Prevents |
|---|---|---|
| **Cryptographic** | SHA256 commit-reveal | Front-running, copying |
| **Physical** | OCR proof of challenge hash | Data fabrication |
| **Statistical** | Gaussian consensus (z > 2σ = 0) | Outliers, collusion |

### Reward Function:
```
Score = 60% Accuracy + 30% Proof + 10% Speed
```

Honest miners who physically observe and report quickly → **maximum emission**.

---

## SLIDE 5: MARKET OPPORTUNITY

# 📊 A $2.2 Billion Addressable Market

### The Crisis: Inflation data across the Cono Sur is broken

Argentina's INDEC scandal (2007-2015) understated inflation by **2-3×**. Venezuela's hyperinflation made official data meaningless. Even in Chile and Brazil, CPI is **monthly, urban-only, and 2-4 weeks delayed**. Over 700M Latin Americans make economic decisions on stale or falsified data.

### The Opportunity: Remittances + Fintech

- **$150B/year** in remittances to LatAm — senders cannot verify purchasing power
- **300M+** fintech users (Nubank, Ualá, Mercado Pago) need real inflation dashboards
- **Zero existing oracles** serve physically-verified consumer price data

| Segment | Market Size | Use Case |
|---|---|---|
| Inflation-linked bonds | $800B in LatAm | Real-time CPI alternative |
| Remittance platforms | $150B annual flow | Purchasing power transparency |
| Fintech apps | 300M+ users | Real inflation dashboards |
| Insurance & agriculture | $50B premiums | Price + weather models |
| Central banks & NGOs | — | Independent data verification |
| DeFi / stablecoins | Growing | On-chain LatAm CPI feeds |

### No existing oracle serves this market.
- Chainlink: Financial prices, no consumer goods
- Pyth: DeFi-focused, no physical verification
- Government: Delayed, potentially manipulated

---

## SLIDE 6: COMPETITIVE ADVANTAGE

# 🏆 Why LatAm Truth Oracle Wins

| Feature | OraTruth | Chainlink | Gov Stats | Web Scraping |
|---|---|---|---|---|
| Real-time | ✅ Every 2 min | ✅ | ❌ Monthly | ✅ |
| Physical verification | ✅ OCR proof | ❌ | ❌ | ❌ |
| Consumer goods | ✅ | ❌ | ✅ | Partial |
| Censorship resistant | ✅ | Partial | ❌ | ❌ |
| LatAm coverage | ✅ 5 countries | ❌ | Per-country | Limited |
| Anti-manipulation | ✅ Commit-reveal | N/A | ❌ | N/A |

### Our Moat: **Proof of Physical Truth (PoPT)**
We don't trust local reports — we cryptographically **force** miners to prove physical presence. You can fake a JSON, but you can't fake reality through our Commit-Reveal pipeline.

---

## SLIDE 7: BITTENSOR FIT

# 🧠 Why Bittensor Is the Perfect Platform

### 1. Decentralized Incentives
TAO emission naturally incentivizes geographic coverage — miners in underserved regions earn more (less competition).

### 2. Built-In Sybil Resistance
Registration costs + staking prevent cheap identity attacks.

### 3. Composability
Other subnets can consume OraTruth price feeds as inputs.

### 4. Censorship Resistance
No government can shut down a globally distributed miner network.

### 5. Proof of Effort
Physical price observation requires **real-world labor** — the most genuine form of proof of work.

---

## SLIDE 8: GO-TO-MARKET

# 🚀 Growth Strategy

### Phase 1: Prove It Works (Months 1-6)
- 5-10 miners across Chile, Argentina, Brazil
- Demonstrate data quality on testnet
- Publish weekly "LatAm Bread Index" on Twitter/X

### Phase 2: Scale the Network (Months 6-12)
- 50+ miners, 5+ countries
- Partner with fintech companies (Mercado Pago, Nubank)
- Academic partnerships for research papers

### Phase 3: Monetize (Year 2+)
- API access: data consumers pay TAO for real-time feeds
- Enterprise packages for institutional clients
- Licensing to central banks and international organizations

### Bootstrap Incentives:
- **Miners**: Be the only reporter in your city = max emission
- **Validators**: Early quality-setters earn maximum dividends
- **Consumers**: Free during testnet, discounted Year 1

---

## SLIDE 9: TRACTION & ROADMAP

# 📈 Current Status: Working Demo

### ✅ Built & Functional
- **Commit-Reveal Protocol**: Full Pydantic models with hash verification (`protocol/oracle.py`)
- **Gaussian Reward Engine**: 60/30/10 weighted scoring with NumPy (`reward/reward.py`)
- **EasyOCR Proof Verification**: English + Spanish models (`reward/ocr_verifier.py`)
- **Validator + Miner Neurons**: Complete Bittensor integration (`neurons/`)
- **🖥️ Live Streamlit Dashboard** (`dashboard/demo_app.py`): Dark-theme professional UI with miner leaderboard, real-time price oracle feed, epoch history charts, and commit-reveal status panel — **fully functional and demo-ready**
- **🎬 Terminal Simulation** (`scripts/local_demo.py`): Interactive commit-reveal demo with 5 miners (3 honest, 1 copycat, 1 attacker) showing attack detection and reward distribution in real-time
- **Automated Deployment Scripts**: Docker, faucet, miner spawning

### 🎯 What the Demo Proves
- Honest miners earn **~79%** of epoch rewards
- Copycat (front-running) is detected → proof fails → reduced score
- Attacker (price manipulation) is detected → accuracy = 0 → near-zero reward
- The dashboard provides **Bloomberg-grade** monitoring for any fintech consumer

### 🗓️ Roadmap

| Quarter | Milestone |
|---|---|
| **Q1 2026** | Testnet deployment, 5 miners operational |
| **Q2 2026** | Mainnet launch, 20 miners, first API users |
| **Q3 2026** | 50+ miners, fintech partnerships, first revenue |
| **Q4 2026** | 100+ miners, 10+ countries, enterprise API |

---

## SLIDE 10: THE ASK

# 🤝 Join the Oracle Revolution

### We're building the Bloomberg Terminal for LatAm's real economy.

**What we need:**
- ✅ Selection for Round II (Hackathon)
- 🤝 Bittensor ecosystem connections
- 🌎 Miner operators across Latin America
- 💼 Fintech partners for data consumption

**We have the math. We have the code. We are ready for Testnet execution in Round II.**

### Contact
- **GitHub**: [github.com/zzzbedream/OraTruth-Latam](https://github.com/zzzbedream/OraTruth-Latam)
- **Twitter/X**: @LatAmTruthOracle

> *"In a region where data is power, decentralizing that data is an act of liberation."*

---

*LatAm Truth Oracle — Bittensor Subnet Ideathon 2026*

 🏛️ LatAm Truth Oracle — Subnet Design Proposal

## Bittensor Subnet Ideathon — Round I Submission

---

## Executive Summary

**LatAm Truth Oracle** is a Bittensor subnet that creates a decentralized, censorship-resistant network for verifying real-world consumer prices across Latin America. Miners physically observe and report prices of basic goods (bread, milk, rice) from local markets, while validators verify data integrity through a cryptographic commit-reveal protocol and OCR-based proof of physical presence. The subnet produces a trustworthy, real-time price index that no single entity can manipulate.

---

## 1. The Problem

### Why Latin America Needs a Price Oracle

Latin America faces a **crisis of data trust**:

- **Argentina**: The government manipulated official inflation statistics (INDEC scandal, 2007-2015), understating inflation by 2-3x. Citizens created independent indices like PriceStats and Billion Prices Project to track real prices.
- **Venezuela**: Hyperinflation made official data meaningless. The black market exchange rate diverged 1000x from official rates.
- **Chile, Colombia, Brazil**: While more transparent, price data still depends on centralized statistical offices with delayed publication (monthly vs. real-time) and limited geographic coverage.

**The core problem**: Price data in emerging markets is either manipulated, delayed, or geographically sparse. This affects:
- Consumer purchasing decisions
- Inflation-linked financial instruments
- Remittance value estimation (senders don't know what recipients can buy)
- Policy-making and economic research

### Why Existing Solutions Fall Short

| Solution | Limitation |
|---|---|
| Government statistics (INDEC, INE, IBGE) | Delayed (monthly), potentially manipulated, limited coverage |
| Billion Prices Project (MIT) | Scrapes online prices only (doesn't reflect physical market reality) |
| Chainlink/Band Protocol | Financial prices only, no consumer goods, no physical verification |
| Traditional surveys | Expensive, slow, sample-based, not real-time |

**No existing oracle provides physically-verified, real-time consumer price data from on-the-ground observers.** This is the gap LatAm Truth Oracle fills.

---

## 2. Incentive & Mechanism Design

### 2.1 Emission and Reward Logic

The reward function uses a **weighted scoring model** with three components:

```
Final Score = 0.60 × Accuracy + 0.30 × Proof + 0.10 × Latency
```

| Component | Weight | Formula | Purpose |
|---|---|---|---|
| 🎯 **Accuracy** | 60% | `exp(-z²/2)` where z = (price - μ) / σ | Gaussian consensus — miners reporting prices close to the group mean are rewarded. Outliers beyond 2σ receive zero. |
| 📸 **Proof** | 30% | Binary: `PASS = 1.0, FAIL = 0.0` | OCR verification of a physical proof image containing the challenge hash. This proves physical presence at a market. |
| ⚡ **Latency** | 10% | `exp(-t / 10)` | Exponential decay rewarding faster responses. Encourages timely data. |

**Why these weights?**
- Accuracy dominance (60%) ensures the network converges on truthful prices
- Proof requirement (30%) prevents pure data fabrication — miners must physically observe
- Latency bonus (10%) incentivizes freshness without penalizing honest miners in remote areas

### 2.2 Incentive Alignment

**Miners are incentivized to:**
- Report accurate, truthful prices (Gaussian consensus rewards honesty)
- Provide physical proof (30% of reward requires OCR verification)
- Respond quickly (latency bonus)
- Cover diverse geographic areas (different markets = different price signals = less competition)

**Validators are incentivized to:**
- Score honestly (their dividends depend on the quality of miners they stake on)
- Maintain uptime (missing rounds means missing dividend emission)
- Run OCR verification (proof scoring affects miner selection)

### 2.3 Anti-Adversarial Mechanisms

| Attack Vector | Defense Mechanism |
|---|---|
| **Collusion** (miners agree on fake prices) | Commit-reveal protocol prevents seeing others' answers before committing. 60s anti-gaming timeout between commit and reveal. |
| **Fabrication** (making up prices without observing) | OCR proof requirement: miners must photograph a handwritten challenge hash at the market. Without physical presence, no proof = 0 reward for 30% of the score. |
| **Sybil** (creating many identities) | Bittensor's registration cost + stake requirement makes sybil attacks expensive. Gaussian consensus also dilutes sybil miners (many identical prices look suspicious). |
| **Front-running** (copying others' answers) | Commit-reveal protocol: miners submit `SHA256(price_data)` first (blind), then reveal data only after timeout. Impossible to copy. |
| **Replay** (submitting old data) | Each round has a unique challenge hash. Old data produces hash mismatch = 0 score. |
| **Low-effort** (random/garbage data) | Gaussian outlier removal (z > 2σ = 0 accuracy score). Garbage data is statistically far from consensus. |

### 2.4 Proof of Effort / Proof of Intelligence

This subnet qualifies as a **genuine Proof of Intelligence/Effort** — a category of Useful Proof of Work where the labor required to earn emission cannot be replicated by simple bots or automated pipelines:

1. **Physical labor (irreducible cost)**: Miners must physically visit markets and observe prices. This creates an irreducible cost-of-participation that no API call, web scraper, or LLM can replicate. The OCR proof photo must be taken *at the point of sale* with the round's unique challenge hash handwritten and visible — a bot cannot generate a valid image without being physically present.
2. **Domain intelligence**: Reporting accurate prices requires situated knowledge of local markets — understanding regional product names, currency denominations, seasonal variation, and geographic pricing differences. A miner in Santiago knows that "pan de molde" and "marraqueta" are different bread categories with different prices; a bot does not.
3. **Cryptographic commitment (anti-copying)**: The commit-reveal protocol requires miners to submit `SHA256(price_data|nonce)` *before* seeing any other miner's answer. Combined with the 60-second anti-gaming timeout, this makes front-running and data-copying mathematically impossible within a round.
4. **Round-bound proof (anti-replay)**: Each challenge hash is unique per round (~120s). A pre-captured proof image is useless because EasyOCR (English + Spanish) validates the exact challenge string at character level. Old photos produce hash mismatches.
5. **Statistical verification (anti-fabrication)**: Even if a miner bypasses proof somehow, fabricated prices are eliminated by Gaussian consensus — random or extreme values fall beyond 2σ and receive zero accuracy score (60% of the total reward).

> **In summary**: The combination of physical presence requirements, domain-specific intelligence, cryptographic binding, and statistical consensus creates a verification pipeline where no single automation strategy can achieve a competitive score. This is not proof-of-compute — it is **Proof of Physical Truth (PoPT)**: real humans, in real markets, observing real prices.

### 2.5 High-Level Algorithm

```
┌──────────────────────────────────────────────────────────┐
│                    VALIDATOR ROUND                        │
├──────────────────────────────────────────────────────────┤
│                                                          │
│  1. CHALLENGE                                            │
│     └─ Generate unique challenge_hash = SHA256(random)   │
│     └─ Define target: currencies=[CLP,ARS,BRL]           │
│                       products=[bread,milk,rice]         │
│                                                          │
│  2. COMMIT PHASE (30s timeout)                           │
│     └─ Send challenge_hash to all miners                 │
│     └─ Miners respond: commit_hash = SHA256(price_data)  │
│     └─ Collect commit_hash from each miner               │
│                                                          │
│  3. ANTI-GAMING TIMEOUT (60s)                            │
│     └─ No communication allowed                          │
│     └─ Prevents last-second price adjustments            │
│                                                          │
│  4. REVEAL PHASE (30s timeout)                           │
│     └─ Request full data from committed miners           │
│     └─ Miners send: prices + proof_image + nonce         │
│     └─ Collect reveals                                   │
│                                                          │
│  5. VERIFICATION                                         │
│     └─ Verify SHA256(reveal) == commit_hash              │
│     └─ Run OCR on proof_image → check challenge_hash     │
│     └─ Compute Gaussian consensus on prices              │
│                                                          │
│  6. SCORING                                              │
│     └─ accuracy = exp(-z²/2)                      [60%]  │
│     └─ proof    = 1.0 if OCR passes, else 0.0    [30%]  │
│     └─ latency  = exp(-response_time / 10)       [10%]  │
│     └─ final    = 0.6*acc + 0.3*proof + 0.1*lat         │
│                                                          │
│  7. WEIGHT SETTING                                       │
│     └─ Normalize scores → set on-chain weights           │
│     └─ Emission flows to high-scoring miners             │
│                                                          │
│  8. REPEAT (every 120 seconds)                           │
│                                                          │
└──────────────────────────────────────────────────────────┘
```

---

## 3. Miner Design

### 3.1 Miner Tasks

Miners observe and report real-world consumer prices from physical markets in Latin America. Each round, a miner:

1. Receives a challenge hash and target parameters (currencies, products)
2. Observes current prices at their local market
3. Photographs a handwritten note of the challenge hash (proof of presence)
4. Computes SHA256 hash of their data and submits the commit
5. After the timeout, reveals their full data + proof image

### 3.2 Input → Output Format

**Input (from Validator):**
```json
{
  "challenge_hash": "a1b2c3d4...",
  "phase": "commit",
  "target_currencies": ["CLP", "ARS", "BRL"],
  "target_products": ["bread", "milk", "rice"]
}
```

**Output — Commit Phase:**
```json
{
  "commit_hash": "SHA256(price_data|challenge_response|nonce)",
  "miner_hotkey": "5Grw...QY"
}
```

**Output — Reveal Phase:**
```json
{
  "prices": [
    {"currency": "CLP", "product": "bread", "price": 1200.00, "location": "Santiago", "source": "direct"},
    {"currency": "CLP", "product": "milk", "price": 1500.00, "location": "Santiago", "source": "direct"}
  ],
  "proof_image_b64": "<base64 encoded image>",
  "challenge_response": "SHA256(challenge_hash + hotkey)",
  "nonce": "random_hex"
}
```

### 3.3 Performance Dimensions

| Dimension | Measurement | Impact |
|---|---|---|
| **Accuracy** | Distance from Gaussian mean (z-score) | 60% of reward |
| **Proof Quality** | OCR detects challenge hash in image | 30% of reward |
| **Speed** | Response time in seconds | 10% of reward |
| **Coverage** | Unique locations reported | Future dimension (diversity bonus planned) |

### 3.4 Miner Modes

| Mode | Behavior | Expected Outcome |
|---|---|---|
| **Real** | Physical observation with proof photos | High scores across all dimensions |
| **Dummy** | Generates realistic random prices ±5% | Used for testing; moderate accuracy |
| **Noisy** | Random prices with high variance | Low accuracy, demonstrates penalty |
| **Dead** | No response | Zero emission, demonstrates timeout handling |

---

## 4. Validator Design

### 4.1 Scoring and Evaluation Methodology

Validators orchestrate the full commit-reveal cycle and apply a multi-factor scoring function:

1. **Hash Verification** (binary gate): `SHA256(reveal_data) == commit_hash`
   - If mismatch → immediate disqualification (score = 0)

2. **Gaussian Consensus** (accuracy, 60%):
   - Collect all revealed prices for each (currency, product) pair
   - Compute μ (mean) and σ (standard deviation) across miners
   - Each miner's z-score: `z = |price - μ| / σ`
   - Score: `exp(-z²/2)` — bell curve centered on consensus
   - Outliers beyond 2σ receive exactly 0

3. **OCR Proof Verification** (proof, 30%):
   - Decode base64 image from miner
   - Run EasyOCR (English + Spanish models)
   - Check if first 8+ characters of challenge hash appear in detected text
   - Binary: PASS (1.0) or FAIL (0.0)

4. **Response Time** (latency, 10%):
   - Measured from query sent to response received
   - Score: `exp(-time / 10)` — exponential decay with 10s constant
   - 1 second response → 0.90 score
   - 5 second response → 0.61 score
   - 30 second response → 0.05 score

### 4.2 Evaluation Cadence

- **Round duration**: ~120 seconds (commit + timeout + reveal + scoring)
- **Rounds per hour**: ~30
- **Weight setting**: Every round (on-chain)
- **Metagraph sync**: Every round (to detect new/departing miners)

### 4.3 Validator Incentive Alignment

Validators earn dividends proportional to the quality of miners they help select. By scoring miners accurately:
- High-quality miners receive more emission → produce better data
- The network's data quality improves → attracts consumers/buyers
- Increased demand → TAO price appreciation → validator profit

Lazy validators who don't verify proofs or compute consensus correctly will:
- Stake on poor-quality miners
- Lose dividends relative to competing validators
- Be naturally de-selected by the network

---

## 5. Business Logic & Market Rationale

### 5.1 The Problem We Solve

**Real-time, physically-verified consumer price data for Latin America.**

Current alternatives are either manipulated (government stats), delayed (monthly reports), online-only (web scraping misses physical markets), or non-existent for consumer goods (crypto oracles cover financial assets).

#### The Crisis of Inflation Data in the Cono Sur

Latin America's Southern Cone (Argentina, Chile, Uruguay, Brazil) has endured decades of inflation volatility, institutional distrust, and outright data manipulation:

- **Argentina (INDEC scandal, 2007-2015)**: The government systematically understated inflation by 2-3×, leading the IMF to issue a "declaration of censure" — the first against a G20 nation. Citizens responded by creating grassroots price-tracking indices (PriceStats, Billion Prices Project), proving the demand for independent, verifiable price data.
- **Venezuela**: Hyperinflation exceeding 1,000,000% rendered official statistics meaningless. The black-market "DolarToday" rate became the de facto price reference — a decentralized oracle born of necessity.
- **Chile & Brazil**: While more transparent, official CPI data is published monthly with a 2-4 week lag, covers limited product baskets, and samples only major urban centers — missing the reality of rural and peri-urban markets.

The consequence: over 700 million Latin Americans make economic decisions based on data that is stale, sparse, or actively falsified.

#### The Remittance Corridor Opportunity

The LatAm remittance market exceeds **$150 billion annually** (World Bank, 2024), representing the largest source of foreign income for many families. Yet senders in the US, Spain, and Europe have no way to verify the purchasing power of the money they transfer. *"I sent $200 to my mother in Buenos Aires — but can she actually buy a week of groceries?"* No existing fintech, oracle, or government API answers this question with physically-verified, real-time data.

OraTruth fills this gap by producing **ground-truth consumer prices** from on-the-ground observers — data that remittance platforms (Wise, Remitly, Mercado Pago) can integrate via API to show senders the real-world value of their transfers. This transforms remittances from a blind transfer into an informed financial decision.

### 5.2 Competing Solutions

| Competitor | Type | Limitation vs. OraTruth |
|---|---|---|
| **INDEC/INE/IBGE** | Government | Slow (monthly), potentially manipulated, limited coverage |
| **Billion Prices Project** | Academic | Online-only, no physical verification |
| **Chainlink** | Crypto oracle | Financial prices only, no consumer goods |
| **Pyth Network** | Crypto oracle | DeFi-focused, no real-world verification |
| **RedAtam (CEPAL)** | Regional statistics | Census data, years delayed |
| **None within Bittensor** | — | No existing subnet targets LatAm price data |

### 5.3 Why Bittensor Is the Right Platform

1. **Decentralized verification**: No single entity controls the data pipeline. Validators independently verify miner submissions using the same mathematical consensus.

2. **Economic incentives**: TAO emission naturally incentivizes geographic coverage. As miners in unprofitable regions join (less competition), they earn more → filling data gaps.

3. **Censorship resistance**: Governments cannot shut down the network. Even if one country bans participation, miners in other countries continue reporting.

4. **Built-in Sybil resistance**: Registration costs and staking requirements prevent cheap identity flooding.

5. **Composability**: Other subnets can consume OraTruth's price feeds as inputs to their own models (e.g., DeFi protocols, prediction markets, fair trade verification).

### 5.4 Path to Long-Term Adoption

**Phase 1 (0-6 months)**: Testnet proof of concept with 5-10 miners across Chile, Argentina, and Brazil. Demonstrate data quality and mechanism soundness.

**Phase 2 (6-12 months)**: Mainnet launch with 50+ miners. Partner with:
- Fintech companies serving LatAm remittance corridors
- Academic researchers studying inflation
- NGOs monitoring food security

**Phase 3 (12-24 months)**: API commercialization. Sell real-time price feeds to:
- Inflation-linked bond traders
- Insurance companies (crop/weather + price models)
- Central banks seeking independent verification
- Stablecoin protocols needing local CPI data

**Revenue model**: Data consumers pay TAO to access the price feed, which flows to subnet validators and miners.

---

## 6. Go-To-Market Strategy

### 6.1 Initial Target Users

| User Segment | Use Case | Why They Care |
|---|---|---|
| **LatAm fintech apps** (Mercado Pago, Ualá, Nubank) | Real-time purchasing power data | Helps users understand real inflation |
| **Remittance platforms** (Wise, Remitly) | "What can $100 USD actually buy in Chile?" | Value proposition for senders |
| **Academic researchers** | Independent inflation tracking | Alternative to government data |
| **DeFi protocols** | LatAm stablecoin price feeds | No existing oracle covers this |

### 6.2 Distribution & Growth Channels

1. **Bittensor ecosystem**: Cross-promote with other LatAm-focused subnets. Publish on Bittensor Discord and forums.
2. **Crypto Twitter/X**: Weekly data reports showing "What does bread cost across LatAm today?" with visualizations.
3. **Academic partnerships**: Share data with university economics departments for research papers (free advertising).
4. **Developer ecosystem**: Open-source the miner client so anyone in LatAm can contribute from their phone.

### 6.3 Incentives for Early Participation

| Role | Bootstrap Incentive |
|---|---|
| **Miners** | Low competition in early rounds = high emission per miner. Geographic first-mover advantage (be the only miner in Bogotá = guaranteed rewards). |
| **Validators** | Early validators set the quality standard and earn maximum dividends before competition increases. |
| **Data consumers** | Free access during testnet phase. Discounted API during first 6 months of mainnet. |

---

## 7. Architecture

```
┌─────────────────────────────────────────────────────────────┐
│                    BITTENSOR NETWORK                         │
│                                                             │
│  ┌─────────────────┐     ┌──────────────────────────────┐  │
│  │    VALIDATORS    │     │         MINER FLEET          │  │
│  │                  │     │                              │  │
│  │ ┌─────────────┐  │     │  ⛏️ Santiago (CLP)           │  │
│  │ │ Challenge    │──┼────▶│  ⛏️ Buenos Aires (ARS)      │  │
│  │ │ Generator    │  │     │  ⛏️ São Paulo (BRL)         │  │
│  │ └──────┬──────┘  │     │  ⛏️ CDMX (MXN)             │  │
│  │        │         │     │  ⛏️ Bogotá (COP)            │  │
│  │ ┌──────▼──────┐  │     │                              │  │
│  │ │ Commit-     │  │     │  Input: Challenge hash       │  │
│  │ │ Reveal      │◀─┼─────│  Output: Prices + Proof IMG  │  │
│  │ │ Verifier    │  │     │                              │  │
│  │ └──────┬──────┘  │     └──────────────────────────────┘  │
│  │        │         │                                       │
│  │ ┌──────▼──────┐  │                                       │
│  │ │ OCR Proof   │  │     ┌──────────────────────────────┐  │
│  │ │ Verifier    │  │     │      DATA CONSUMERS          │  │
│  │ │ (EasyOCR)   │  │     │                              │  │
│  │ └──────┬──────┘  │     │  📊 Fintech APIs             │  │
│  │        │         │     │  📈 DeFi Protocols            │  │
│  │ ┌──────▼──────┐  │     │  🔬 Research Institutions    │  │
│  │ │ Gaussian    │  │     │  💱 Remittance Platforms      │  │
│  │ │ Scorer      │──┼────▶│                              │  │
│  │ │ (60/30/10)  │  │     └──────────────────────────────┘  │
│  │ └──────┬──────┘  │                                       │
│  │        │         │                                       │
│  │ ┌──────▼──────┐  │                                       │
│  │ │ Weight      │  │                                       │
│  │ │ Setter      │  │                                       │
│  │ └─────────────┘  │                                       │
│  └─────────────────┘                                        │
└─────────────────────────────────────────────────────────────┘
```

---

## 8. Technical Stack

| Component | Technology | Purpose |
|---|---|---|
| Network | Bittensor SDK 10.1.0 | Subnet registration, metagraph, weight setting |
| Protocol | Pydantic + btSynapse | Typed data models for commit/reveal |
| Consensus | NumPy (Gaussian) | Statistical price verification |
| Proof | EasyOCR (EN+ES) | Optical character recognition of proof images |
| Monitoring | Streamlit | Real-time dashboard with metagraph viewer |
| Console | Rich | Premium CLI output with tables and progress bars |
| Hashing | SHA256 (hashlib) | Commit-reveal integrity |
| Container | Docker (optional) | Reproducible deployment |

---

## 9. Repository Structure

```
OraTruth/
├── protocol/
│   ├── __init__.py
│   └── oracle.py          # OracleSynapse, PriceData, commit-reveal models
├── reward/
│   ├── __init__.py
│   ├── reward.py           # Gaussian scoring (60/30/10)
│   └── ocr_verifier.py     # EasyOCR proof verification
├── neurons/
│   ├── validator.py        # Async commit-reveal orchestrator
│   └── miner.py            # Price reporter (dummy/real/noisy/dead modes)
├── utils/
│   ├── __init__.py
│   ├── crypto.py           # SHA256 hashing, nonce generation
│   └── logger.py           # Rich console output
├── dashboard/
│   └── app.py              # Streamlit live monitoring UI
├── scripts/
│   ├── auto_faucet.sh      # Automated TAO acquisition
│   ├── spawn_miners.sh     # Dummy miner fleet for testing
│   └── deploy_testnet.sh   # Full testnet deployment automation
├── docs/
│   └── SUBNET_DESIGN_PROPOSAL.md  # ← This document
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
└── README.md
```

---

*LatAm Truth Oracle — Verifiable Real-World Price Data for Latin America*
*Bittensor Subnet Ideathon — Round I Submission — February 2026*

Completed Round I: Full system architecture designed, Python neurons for Validator/Miner implemented, local simulation successful, and real-time monitoring dashboard functional. Ready for Testnet execution in Round II

"Bootstrapped / Seeking Round II Grant."