Abstract
PREDIBO is an AI-powered protocol for automated trading on prediction markets — a sector that recorded over $44 billion in trading volume in 2025 and is rated by analysts as one of the fastest-growing segments of financial markets. The platform combines advanced machine learning models with real-time data analysis to systematically identify and trade markets with positive expected value on platforms like Polymarket ($21.5B volume in 2025, valuation $9–12B).
At the heart of the platform is PrediBo — the AI-powered trading bot that fully automatically analyses, scores and executes trades. Users connect their Polymarket account via a secure API integration, define their interests and risk parameters, and hand over analysis and execution to PrediBo.
The MVP is complete and live in production at app.predibo.com. The first 50 PREDIBO token holders can apply as beta testers and receive direct access to the running system — before the official launch.
The platform is monetised via a SaaS subscription model and complemented by the PREDIBO Utility Token — an SPL Token on Solana. Token holders receive 25% of net subscription revenue monthly, proportional to tokens held. Total supply is fixed at 100,000,000 PREDIBO. Hard cap: $8,000,000.
Introduction & Problem Statement
2.1 The Rise of Prediction Markets
Prediction markets have evolved within a few years from a niche product into one of the fastest-growing asset classes worldwide. Industry volume reached over $44 billion in 2025 — a 130-fold increase versus early 2024. Industry analysts project further revenue growth to over $10 billion annually by 2030.
2.2 The Problem
- Information overload — Hundreds of active markets simultaneously. Without systematic analysis it is nearly impossible to consistently identify mispricings.
- Time requirement — Successful trading requires continuous monitoring and fast reaction to news — around the clock, 7 days a week.
- Missing tools — Almost no specialised tools for prediction market analysis exist for retail users.
- Emotional decisions — Cognitive biases (confirmation bias, overconfidence, loss aversion) lead to suboptimal decisions.
2.3 Market Opportunity
| Platform | Trading Volume 2025 | Status |
|---|---|---|
| Polymarket | $21.5B | Live |
| Kalshi | $8.2B | Month 10+ |
| Further markets | — | Roadmap |
The Solution: PREDIBO Protocol
3.1 Vision
PREDIBO makes systematic, data-driven trading on prediction markets accessible to everyone — regardless of technical knowledge, time resources or data analysis skills.
3.2 PrediBo Bot
PrediBo is the AI-powered, fully automated trading bot at the heart of the platform. It continuously analyses active markets, calculates its own probabilities based on a multi-tier signal model, compares these against the market price, and automatically places positions wherever a statistically significant edge exists — free of emotional decisions, without sleep breaks, fully traceable in the bot log.
3.3 MVP Features Live Today
- Complete user onboarding with KYC
- PrediBo Bot Engine with automated trading on Polymarket
- Live dashboard with portfolio, P&L, and bot log
- Take-profit, stop-loss, and trailing stop exit logic
- Standalone AI reversal signal for exits
- Fast Markets Mode for short-duration markets (<2h)
- Admin panel with user management and server monitoring
Platform & Product Architecture
4.1 User Onboarding
Sign up at app.predibo.com with email and KYC verification.
Connect a Solana wallet (Phantom, Solflare, Backpack).
Import your Polymarket private key and proxy wallet address securely. Keys are encrypted with AES-256 and never stored in plaintext.
Set risk profile (Conservative / Moderate / Aggressive), budget limits, preferred market categories, and exit logic.
PrediBo immediately begins market analysis and trading.
4.2 Dashboard Modules
| Module | Description |
|---|---|
| Portfolio Overview | Active and closed positions, total performance, ROI, win rate, P&L curve. CSV export. |
| Bot Activity Log | All trades with signal score, probability vs. market price, position size, confidence. |
| Risk Monitor | Real-time budget utilisation, active exposure, drawdown warnings. |
| Token & Rewards | PREDIBO balance, staking status, upcoming reward distributions. |
PrediBo Engine – How the Bot Works
PrediBo runs automatically every 15 minutes. Each cycle it scans active Polymarket markets, scores them, and places trades on the best opportunities — within the limits configured in Bot Settings.
5.1 The 6-Step Bot Cycle
Reviews all open trades for exit conditions. Qualifying positions are sold before new trades are evaluated.
Fetches all non-expired Polymarket markets matching the user's configured categories.
Checks whether a fresh AI prediction exists (cached up to 6 hours). If not, generates one using live news search and LLM analysis.
Scores each market using the AI prediction (Mode ①) or fallback rule-based model (Mode ②). Only markets above the risk threshold generate a trade signal.
Verifies daily, weekly, and total budget remaining; checks rate limits; confirms no open position for this market.
Executes a Fill-or-Kill (FOK) order directly on the Polymarket CLOB. Either fully executed immediately — or discarded. No partial fills, no slippage surprises.
5.2 Fast Markets Mode
- Normal bot: runs every 15 min, skips markets with <2 days remaining
- Fast bot: runs every 2 min, targets short-duration markets specifically
- Skip rate: configurable — what % of normal markets to skip when fast mode is active
5.3 Risk Profiles
| Profile | Min. Score | Trade Size |
|---|---|---|
| Conservative | ≥ 75 / 100 | 5% of daily budget |
| Moderate | ≥ 60 / 100 | 10% of daily budget |
| Aggressive | ≥ 45 / 100 | 20% of daily budget |
Signal Scoring – Two Modes
Used when fresh AI prediction available (<6h old)
| Factor | Weight |
|---|---|
| AI Prediction Score | 55% |
| Volume | 25% |
| Liquidity | 15% |
| Time | 5% |
Used when no fresh AI prediction available
| Factor | Weight |
|---|---|
| Mispricing | 45% |
| Volume | 20% |
| Liquidity | 20% |
| Time | 15% |
Trade Direction Examples
Minimum required edge: 5%. Below that, the bot falls back to Mode ②.
Exit Logic – When Does the Bot Sell?
At the start of every bot cycle, all open positions are reviewed. Exit decisions are evaluated in four tiers — the first condition that triggers causes the position to be sold.
User-Configured Thresholds
| Setting | Trigger |
|---|---|
| Take Profit | Sell when gain ≥ X% from entry price |
| Stop Loss | Sell when loss ≥ X% from entry price |
| Max. Duration | Sell when position is older than the configured time limit |
All three are optional and independent. Leaving a setting blank disables that threshold.
Trailing Stop (with AI Override)
Tracks the highest price the position has reached since entry. If the price drops X% from that peak, the bot evaluates an exit.
AI override only applies to predictions less than 6 hours old.
Always-Active Market Logic
| Condition | Action | Reason |
|---|---|---|
| Price ≤ 5¢ | Sell immediately | Market almost certainly losing — recover residual value |
| Expiry <2h + loss >10% | Sell before resolution | Avoid waiting for a zero payout |
| Loss >40% + held >30min | Dynamic stop-loss | Prevents runaway losses even without stop-loss configured |
Standalone AI Reversal Signal
If the AI produces a fresh, high-confidence estimate that strongly contradicts an open position, the bot exits — even if none of the above conditions triggered.
- Prediction must be less than 2 hours old
- AI confidence must be ≥ 65%
- AI edge against the position must be > 15%
AI Predictions Engine
8.1 What the AI Produces
| Output | Description |
|---|---|
| Estimated Probability | AI's YES probability (0–100%). Compared against market price to find edge. |
| Confidence | How certain the AI is (0–100%). Low confidence = smaller weight in scoring. |
| Sentiment | Bullish / Bearish / Neutral based on recent news analysis. |
| Reasoning | Short explanation of the estimate. Visible in Bot Live Feed and AI Predictions page. |
8.2 Caching & Freshness
- Predictions cached for up to 6 hours to avoid excessive API calls
- For exit decisions, only predictions younger than 2 hours are used
- For the standalone AI reversal signal, freshness requirement: <2h old, confidence ≥ 65%
8.3 AI Predictions Page
The AI Predictions page shows all current market predictions with probability estimates, sentiment, confidence, and reasoning. Users can also manually trigger trades directly from signals shown there.
Technical Infrastructure
9.1 Tech Stack
| Layer | Technology |
|---|---|
| Backend | FastAPI (Python), Uvicorn, SQLAlchemy (async) |
| Database | PostgreSQL 16, Redis 7 (cache & sessions) |
| Scheduler | APScheduler (15-min cycle + 2-min Fast Markets) |
| Polymarket Client | py-clob-client (Gamma API + CLOB API) |
| Frontend | Next.js 14, React 18, TypeScript, Tailwind CSS |
| Cryptography | Fernet AES-256 (API keys), bcrypt (passwords), JWT HS256 |
| Infrastructure | Docker Compose, Nginx, Hetzner Cloud, GitHub Actions |
| Blockchain | Solana (SPL Token, Anchor Framework) |
9.2 System Architecture
9.3 Data Security
- Private keys stored with AES-256 (Fernet) — decrypted only in memory at trade signing
- No employee has access to decrypted credentials
- CORS restricted to predibo.com, app.predibo.com, token.predibo.com
- Multi-sig treasury (Squads Protocol, 3-of-5 signatures)
- Timelock 48h for critical smart contract changes
Tokenomics & PREDIBO Token
10.1 Token Overview
| Parameter | Value |
|---|---|
| Token Name | PREDIBO |
| Standard | SPL Token · Solana |
| Total Supply | 100,000,000 · Fixed |
| Inflation | None |
| Deflation | 25% Burn on token payments + quarterly buybacks |
| Base Price | $0.10695 |
| FDV at Base Price | $10,695,000 |
| Token Sale | token.predibo.com |
10.2 Token Allocation
10.3 Token Utility
Revenue Sharing
25% of net subscription revenue distributed monthly to all token holders — proportional to tokens held. Automated on-chain via Solana Smart Contract.
- Platform Access — stake tokens instead of monthly subscription
- Discounts — tiered discounts on all subscription plans
- Burn — 25% of token payments permanently burned
- Quarterly Buyback & Burn from platform revenue
Token Sale & Financing
11.1 Sale Overview
| Parameter | Value |
|---|---|
| Token Sale URL | token.predibo.com |
| Hard Cap | $8,000,000 |
| Accepted Currencies | EUR · USD · USDC · USDT · SOL |
| Public Sale Tokens | 77,000,000 PREDIBO |
| Base Price | $0.10695 per PREDIBO |
| Month 1 Early Bird | $0.09091 (–15%) |
| KYC / AML | Required for all participants |
| Excluded Jurisdictions | USA, China, others per regulatory requirements |
11.2 Use of Proceeds ($8M)
Revenue-Sharing Model
PREDIBO does not operate its own platform trading account. Profit participation is based exclusively on subscription revenue minus operating costs — predictable, transparent, independent of speculative trading results.
12.1 Calculation Examples
Starter Subscription
| Subscription revenue | $99 / month |
| Cost factor | 50% |
| Net revenue | $49.50 |
| Profit pool (25%) | $12.37 per user |
| At 1,000 users | $12,370 / month |
Pro Subscription
| Subscription revenue | $999 / year |
| Monthly equivalent | $83.25 |
| Net revenue | $41.62 |
| Profit pool (25%) | $10.41 per user/mo |
Subscription Models & Platform Access
| Feature | Starter | Pro | Lifetime |
|---|---|---|---|
| Price | $99 / month | $999 / year | $4,999 one-time |
| PrediBo activation | ✓ | ✓ | ✓ |
| Market categories | 3 | Unlimited | Unlimited |
| Risk profiles | Conservative | All | All + Custom |
| Dashboard | Basic | Extended | Premium |
| Fast Markets | — | ✓ | ✓ |
| API access | — | — | ✓ |
| Support | Priority Email | VIP (Chat + Call) |
DCF Valuation
Monthly FCFs (MRR × 75.5% − Fixed Costs − Marketing) discounted with risk-adjusted rates. Terminal Value based on ARR multiple. EV = Σ discounted FCFs + PV of Terminal Value. Current FDV at base price: $10.7M.
14.1 Scenario Results
| Metric | Conservative | Base Case | Optimistic |
|---|---|---|---|
| Enterprise Value | $19.70M | $120.27M | $407.96M |
| Implied Token Price | $0.1970 | $1.2027 | $4.0796 |
| Upside vs. $0.10695 | +84% | +1,025% | +3,715% |
| Breakeven Month | — | M16 | — |
| ARR Month 36 | $17.25M | $43.88M | $90.67M |
| Terminal Value (PV) | $22.63M | $111.93M | $368.54M |
| Oper. FCF (PV, cum.) | –$2.92M | $8.34M | $39.42M |
14.2 Sensitivity Analysis – Base Case (USD mn)
| Disc.Rate / Multiple | 3× ARR | 5× ARR | 7× ARR | 10× ARR | 12× ARR |
|---|---|---|---|---|---|
| 30% p.a. | $70.26M | $110.20M | $150.14M | $210.05M | $250.00M |
| 35% p.a. | $62.78M | $98.44M | $134.11M | $187.61M | $223.27M |
| 40% p.a. | $56.31M | $88.29M | ★ $120.27M | $168.24M | $200.22M |
| 45% p.a. | $50.69M | $79.48M | $108.26M | $151.44M | $180.22M |
| 50% p.a. | $45.79M | $71.79M | $97.79M | $136.79M | $162.79M |
14.3 Monthly DCF Table – Base Case (M1–M16, excerpt)
| Mo. | Subs | MRR | Rev.Share | Fixed | Marketing | FCF | PV(FCF) | Cum. PV |
|---|---|---|---|---|---|---|---|---|
| M1 | — | $0 | $0 | $85K | $60K | –$145K | –$141K | –$141K |
| M3 | — | $0 | $0 | $85K | $80K | –$165K | –$152K | –$439K |
| M4 | 100 | $9.8K | $1.2K | $95K | $220K | –$307.6K | –$275K | –$714K |
| M6 | 360 | $35.1K | $4.4K | $95K | $280K | –$348.5K | –$294.5K | –$1.31M |
| M9 | 1,100 | $107.3K | $13.4K | $125K | $160K | –$204K | –$158.5K | –$1.85M |
| M12 | 2,250 | $219.4K | $27.4K | $125K | $150K | –$109.4K | –$78.1K | –$2.15M |
| M16 ★ | 5,200 | $507K | $63.4K | $170K | $195K | +$17.8K | +$11.4K | Breakeven |
Security & Audits
- Smart Contract Audits: Min. two independent firms before token launch (e.g. OtterSec, Neodyme, Trail of Bits). Reports publicly accessible. Bug bounty programme.
- API Key Protection: AES-256 Fernet encryption. Private keys never stored in plaintext. Decrypted only in memory at trade signing.
- Auth: JWT HS256, bcrypt for passwords, automatic token refresh. CORS restricted to predibo.com domains.
- MFA for all user accounts. Regular penetration tests. SOC 2 Type II certification as medium-term goal.
- Treasury: Multi-sig wallets (Squads Protocol, 3-of-5 signatures). 48h timelock for critical contract changes. Monthly on-chain transparency reports.
Roadmap
Team & Partners
- CEO / Founder: Vision, strategy, business development
- CTO: Software architecture, FastAPI, Solana smart contracts
- Head of AI/ML: Machine learning, probabilistic modelling, NLP
- Head of Product: UX/UI, Next.js, feature development
- Head of Compliance: Regulatory, KYC/AML, jurisdiction
Core team members with full backgrounds and LinkedIn profiles to follow. Advisors from prediction markets, crypto, quantitative finance, and regulation. Partners: Solana audit firms, market makers, exchange partners.
Risk Factors
- Market risk: The PrediBo Engine can generate losses despite positive signals. Past performance is not indicative of future results.
- Regulatory risk: The regulatory landscape for prediction markets and token sales is continuously evolving.
- Platform risk: Initially dependent on the Polymarket API. Multi-platform strategy reduces this dependency long-term.
- Technology risk: Despite audits, Solana smart contracts may contain vulnerabilities.
- Blockchain risk: Solana network outages or protocol changes can affect operations.
- Liquidity risk: The PREDIBO token can be subject to significant price fluctuations after listing.
- Token classification: Revenue-sharing component could lead to security classification in certain jurisdictions. Legal review is mandatory.
- Staking risk: At average >50M staked tokens, reward requirement exceeds reserve. Guardrail system enables rate adjustment.
Legal Disclaimer
This whitepaper is provided for informational purposes only and does not constitute an offer, solicitation or recommendation to buy or sell tokens or any financial instruments.
The PREDIBO token is a utility token and confers no ownership rights, no dividends, no governance rights and no claims on company assets. The described profit participation is based on subscription revenues minus costs and does not represent a guaranteed return.
Every person participating in the token sale should inform themselves about applicable laws in their jurisdiction and obtain independent legal and financial advice if needed.
Forward-looking statements are based on current assumptions. Actual results may differ materially. Participation in the token sale and use of the platform is at the user's own risk.