1. Abstract
DSRCH (Decentralized Search Ranking Consensus Hypernetwork) is an open, permissionless search engine protocol in which anyone can operate a node, crawl and index the web, serve search queries, and earn DSRCH tokens as compensation for their work. No central entity controls the index, the ranking algorithm, or the search results.
The protocol introduces Proof-of-Index (PoI) — a novel consensus mechanism where nodes prove they have performed useful work (crawling pages, storing index shards, processing queries) and are rewarded accordingly. Validation is performed by randomly selected validators who re-execute sample work to ensure honesty.
DSRCH runs on a purpose-built Cosmos SDK L1 blockchain with 8 specialized modules, uses libp2p for peer-to-peer communication, Bluge for full-text indexing, and CometBFT for consensus with ABCI++ VoteExtensions embedding PoI verification directly into block finalization.
2. Problem Statement
Today's web search is controlled by a handful of corporations. This centralization creates systemic risks:
- Censorship & Bias — A single entity decides what 8+ billion people see. Results can be manipulated for political, commercial, or ideological reasons with no accountability.
- Privacy Violation — Centralized search engines build detailed user profiles from cookies, browsing history, and fingerprinting. Users are the product.
- Monopoly Economics — Webmasters depend on opaque, frequently changing algorithms. A single ranking update can destroy businesses overnight.
- Single Point of Failure — A single outage or regulatory action can take the world's search capability offline.
- No Shared Ownership — Billions of users generate the data, but only shareholders benefit. Node operators, content creators, and users have no stake.
DSRCH solves these problems by making search a public good — owned by the network participants, governed by DAO, and economically aligned to produce high-quality, uncensored results.
3. Vision & Core Principles
A fully decentralized search engine where anyone can run a node, index the web, serve search queries, and earn DSRCH tokens. No central entity controls the results.
3.1 Decentralized Indexing
Thousands of nodes crawl and index the internet simultaneously. No single point of failure or control. The index is sharded across the network using consistent hashing, ensuring redundancy and availability.
3.2 Search Mining
Proof-of-Index mining: earn DSRCH tokens by crawling pages, serving queries, and storing index shards. Unlike Proof-of-Work (wasted computation) or Proof-of-Stake (capital lockup), PoI rewards useful work that directly benefits the network.
3.3 Privacy First
No cookies, no tracking, no user profiles. Optional onion routing for query anonymity. DSRCH will never collect user data — the protocol is designed to make it technically impossible.
4. System Architecture
DSRCH uses a four-layer architecture separating concerns: user-facing gateway, P2P search network, blockchain consensus, and distributed storage.
4.1 P2P Protocol Stack
| Component | Technology |
|---|---|
| Protocol | libp2p (proven by IPFS/Filecoin) |
| DHT | mDNS + Bootstrap peers for node discovery |
| Transport | QUIC + TCP, Noise encryption |
| Discovery | mDNS + Bootstrap + DHT Walk |
| NAT | AutoRelay + Hole Punching |
| Messaging | GossipSub broadcast |
| API | gRPC for inter-node communication, REST/GraphQL for clients |
| Serialization | Protocol Buffers (Protobuf) for efficient message encoding |
| Storage | RocksDB for node state & metadata (LSM-tree, Facebook) |
5. Crawler Engine
Each Full or Validator node runs a crawler that discovers and indexes web pages. The crawler is designed for politeness, efficiency, and quality.
| Component | Implementation |
|---|---|
| URL Frontier | Mercator-style priority queue, Bloom Filter deduplication |
| Fetcher | Async HTTP/2, robots.txt compliance, per-domain rate limiting |
| Parser | HTML → DOM → text extraction, OG tags, hreflang |
| Index Writer | Bluge inverted index, Zstandard compression |
| Deduplication | SimHash near-duplicate detection (threshold: 0.95) |
| Max Depth | Configurable (default: 5 levels from seed URLs) |
Crawled pages are processed through the quality pipeline (Section 10) before being added to the index. Each node stores a subset of the global index (index shards), determined by consistent hashing of URL domains.
6. Search Query Pipeline
A search query flows through the decentralized network in under 200ms using a 5-stage pipeline:
- Query Parsing — Tokenization, stemming, stop-word removal, spell check, query expansion.
- Shard Routing — Consistent hash determines which index shards contain relevant terms.
- Distributed Search — Scatter-gather: parallel search across multiple shard nodes.
- Scoring (BM25) — Multi-signal scoring combining BM25, PageRank, Freshness, DomainTrust, and SpamScore.
- Merge & Re-rank — Top-K merge sort, optional ML re-ranker (ONNX), LRU cache.
6.1 DSRCH-Rank Formula
All weights are DAO-adjustable via governance proposals. This ensures the community — not a corporation — controls what "relevance" means.
7. Proof-of-Index Consensus
DSRCH introduces Proof-of-Index (PoI) — a unique consensus mechanism where nodes prove useful work: crawling, storing, and querying. Unlike PoW (wasted energy) or pure PoS (capital-only), PoI rewards nodes for actual contributions to search quality.
7.1 Mining Process
- Work — Node crawls pages, serves queries, stores index shards.
- Report — Generates a Work Report with Merkle proof each epoch (1 hour).
- Submit — Submits Work Report as a transaction to the L1 chain.
- Validate — 3 randomly selected validators re-crawl sample URLs and re-execute sample queries.
- Reward — If ≥2/3 validators confirm → DSRCH is minted to node wallet.
7.2 Reward Distribution per Epoch
| Activity | Share |
|---|---|
| Query Processing | 35% |
| Web Crawling | 30% |
| Index Storage | 20% |
| Validation | 15% |
7.3 Reward Formula
The minimum 0.5 multiplier guarantees that even nodes without stake receive 50% of the base reward for useful work. The logarithmic component ensures diminishing returns for large stakes, preventing plutocratic domination.
8. Node Types & Requirements
Choose your role in the network based on your hardware and desired rewards:
| Tier | Role | CPU | RAM | SSD | Bandwidth | Stake | ~Reward/day |
|---|---|---|---|---|---|---|---|
| Light | Search Node | 1 core | 2 GB | 20 GB | 10 Mbps | 1,000 DSRCH | ~15 DSRCH |
| Full | Crawler + Search | 2 cores | 4 GB | 80 GB | 25 Mbps | 5,000 DSRCH | ~50 DSRCH |
| Heavy | Full + Validator | 4+ cores | 8+ GB | 200+ GB | 50+ Mbps | 25,000 DSRCH | ~120 DSRCH |
8.1 Node CLI
# Initialize node
$ dsrch init --mode full --stake 5000
# Start
$ dsrch start --port 9090
# Check status
$ dsrch status
Node ID: 12D3KooWBnxt...
Mode: Full (Crawler + Search)
Peers: 1,247 | Shards: 12 (14.2 GB)
Pages today: 45,230 | Queries: 12,891
Pending: 23.5 DSRCH | Uptime: 99.7%
# Claim rewards
$ dsrch rewards claim9. DSRCH Tokenomics
DSRCH is the native L1 coin of the Cosmos SDK blockchain powering the decentralized search economy.
- Max Supply: 1,000,000,000 DSRCH
- Type: Native L1 Coin (Cosmos SDK)
- Denomination: 1 DSRCH = 1,000,000 udsrch
9.1 Token Allocation
| Category | % | Amount | Details |
|---|---|---|---|
| Mining Rewards | 45% | 450M | 10-year emission, smooth ~20%/yr decay |
| Ecosystem Fund | 20% | 200M | DAO-controlled treasury |
| Team & Advisors | 15% | 150M | 1-year cliff + 3-year linear vest |
| Public Sale | 10% | 100M | 20% TGE, 80% over 12 months |
| Community | 5% | 50M | Airdrops, bounties |
| Liquidity | 5% | 50M | Locked LP for 2 years |
9.2 Emission Schedule (Smooth Decay)
| Year | Reward/Epoch | Annual Emission | Decay |
|---|---|---|---|
| 1 | 100 DSRCH/node | ~112.5M | — |
| 2 | 80 DSRCH | ~90.0M | -20% |
| 3 | 64 DSRCH | ~72.0M | -20% |
| 4 | 51 DSRCH | ~57.4M | -20% |
| 5 | 41 DSRCH | ~46.1M | -20% |
| 6–9 | 33→17 DSRCH | ~37→19M | -20% |
| 10+ | 5 DSRCH (floor) | ~5.6M | min |
9.3 Staking & Slashing
| Staking Parameters | |
|---|---|
| Min Stake | 1,000 DSRCH (Light node) |
| Bonus | Logarithmic scale: more stake → more work → more rewards |
| Delegation | Delegate to any node, earn share of rewards |
| Unbonding | 7 days cooldown period |
| Violation | Penalty |
|---|---|
| Fake Work Report | −10% stake |
| Result Censorship | −20% stake |
| Index Poisoning (stage 1) | −10% stake + jailing |
| Index Poisoning (stage 2 — DAO) | up to −50% stake by DAO vote |
| Excessive Downtime | −1% stake |
| Double Report | −5% stake |
10. Quality Control & Anti-Spam
Multi-layer defense ensuring high-quality search results and preventing index poisoning.
Layer 1 — Automated Crawl-Time Filters
Content Quality Signals
- Text-to-HTML ratio analysis (reject < 10%)
- Keyword stuffing detection (abnormal term density)
- Duplicate content via SimHash (near-duplicate threshold: 0.95)
- Thin content filter (< 100 words of meaningful text)
- Language detection + gibberish classifier
- Hidden text / cloaking detection (render vs. source comparison)
Technical Signals
- robots.txt compliance, SSL/TLS validation
- Page load time measurement (penalize > 5s)
- Mobile-friendliness, redirect chain (max 3 hops), HTTP status codes
Layer 2 — ML-Based Spam Detection
- ONNX Runtime model trained on labeled spam/quality dataset
- Features: term distribution, link patterns, content structure
- Confidence ≥ 0.85 → auto-reject; 0.5–0.85 → manual review
- Link farm / PBN detection via IP/hosting clustering
- Anchor text diversity analysis, reciprocal link schemes
Layer 3 — Decentralized PageRank
- Each node computes local PageRank for its crawled subgraph
- Global PageRank via MapReduce protocol between nodes (every 24h)
- DomainTrust = aggregated PageRank + domain age (logarithmic)
- New domains: trust 0.1, sandbox period 30 days
Layer 4 — Community-Driven Moderation
- Users can flag low-quality/spam results; reports weighted by reporter's DSRCH stake
- 5+ unique reports → site enters review queue; false reporting penalized
- DAO blacklist proposals (simple majority vote), appeals process available
Layer 5 — Economic Incentives
Aligned incentives prevent abuse:
- Node Operators: Two-stage slashing for indexing spam (−10% + jailing; up to −50% by DAO vote)
- Webmasters: Priority indexing costs DSRCH; spam sites get zero visibility
- Validators: Rewarded for catching invalid Work Reports; penalized for approving spam
Quality Score Formula
Pages with QualityScore < 0.3 are excluded. Score 0.3–0.5 = deprioritized. Score > 0.5 = normal ranking.
11. Security Model
| Attack Vector | Defense |
|---|---|
| Sybil Attack | Required staking + PoW on registration |
| Index Poisoning | Challenge verification + ML spam detector |
| Query Manipulation | Multi-node consensus (3+ nodes per query) |
| Eclipse Attack | Min 20 peers + diverse selection |
| DDoS | Rate limiting + PoW challenge |
| SEO Spam | Decentralized PageRank + blacklists |
| Censorship | Cross-node comparison + slashing (−20% stake) |
11.1 Privacy Features
- Query Anonymity: Optional onion routing
- No Tracking: Zero cookies, fingerprinting, or user profiles
- Encryption: TLS for all client-node traffic
- Private Search: ZK-proofs planned for Phase 3+
12. Webmaster Integration
How website owners add their sites to the DSRCH index:
12.1 Discovery Methods
- Automatic Discovery — Crawlers follow links; linked sites are found automatically.
- Submit URL API — Direct submission via
POST /api/v1/submit. - Sitemap Protocol — Full-site indexing via standard sitemap.xml.
- Site Verification — DNS TXT record or HTML meta tag for Webmaster Dashboard access.
12.2 Priority Indexing (Paid)
| Tier | Cost | Crawl Time | Recrawl |
|---|---|---|---|
| Free | 0 DSRCH | 2–24 hours | Weekly |
| Priority | 10 DSRCH | < 1 hour | Daily |
| Premium | 100 DSRCH/mo | < 15 min | Hourly |
Revenue goes to the nodes performing the crawl — creating economic incentive for fast indexing.
13. L1 Chain Architecture
Custom Cosmos SDK blockchain with 8 specialized modules and PoI consensus via ABCI++ VoteExtensions.
| Module | Purpose |
|---|---|
x/noderegistry | Node registration, staking, metadata, delegation |
x/crawlproof | Crawl reports, URL proofs, Merkle verification |
x/queryproof | Query reports, latency proofs, result hashes |
x/indexregistry | Shard registry, replica tracking, sync status |
x/rewards | Epoch rewards, smooth decay, pool distribution |
x/slashing | Penalty logic, jailing, dispute resolution |
x/gov | DAO governance, proposals, voting |
x/srchtoken | Native token, mint/burn, denominations |
13.1 Key Features
- CometBFT Consensus — BFT with ABCI++ VoteExtensions, ~1,000–10,000 TPS.
- EVM Compatibility — Cosmos EVM module enables MetaMask, Solidity contracts, and DeFi deployment.
- IBC Cross-Chain — Native IBC to Osmosis (DEX), Noble (USDC), Axelar (ETH/BSC bridge), and 60+ Cosmos chains.
14. DAO Governance
| Parameter | Value |
|---|---|
| Proposal Deposit | 100,000 DSRCH |
| Discussion Period | 7 days |
| Voting Period | 5 days, 67% supermajority required |
| Timelock | 2 days before execution |
The DAO controls ranking weights, slashing parameters, emission schedule, blacklists, and protocol upgrades. All DSRCH token holders with staked tokens can vote, with voting power proportional to stake.
15. Development Roadmap
Phase 0 — Research & Foundation (3 months) ✓ COMPLETE
- P2P protocol research (libp2p + mDNS + Bootstrap)
- FTS engine benchmarks → Bluge selected
- Tokenomics paper & Whitepaper v23.339
- Branding + Landing page + Search UI
Phase 1 — MVP & Testnet (6 months) ✓ COMPLETE
- P2P network — libp2p, GossipSub, mDNS, DHT
- Crawler + Bluge index (BM25 full-text search)
- Scatter-Gather distributed query pipeline
- Node CLI — 25 commands (init, start, search, wallet, stake, gov, testnet, apikey...)
- Token economy — DSRCH minting, staking, slashing, governance
- Multi-node testnet — 3+ nodes, full mesh discovery
Phase 2 — Beta & Incentivized Testnet (4 months) ← CURRENT
- Public testnet deployment
- Advanced DSRCH-Rank ML scoring
- Anti-spam & domain filtering
- Security audit, Bug bounty program
- Documentation & SDK
Phase 3 — Mainnet Launch (3 months)
- Token Generation Event, Mainnet genesis
- Webmaster tools, Polished Web UI v2
- IPFS frontend deployment, API docs + SDK
Phase 4 — Growth (6+ months)
- DAO launch, Mobile app (iOS/Android)
- Browser extension, Image/video search
- AI features (semantic search), Cross-chain bridges
16. Success Metrics (Year 1 Targets)
| Metric | Target |
|---|---|
| Active Nodes | 500+ |
| Indexed Pages | 100M+ |
| Avg Response Time | < 200ms |
| Queries/Day | 10K+ |
| Network Uptime | 99% |
| DEX Liquidity | $1M+ |
17. Conclusion
DSRCH represents a fundamental rethinking of web search. By combining decentralized indexing, Proof-of-Index consensus, and token-aligned incentives, DSRCH creates a search engine that is:
- Uncensorable — No single entity controls what billions of people can find.
- Private — Zero tracking, with optional onion routing for complete anonymity.
- Fair — Transparent ranking formula, DAO-governed, open to all participants.
- Sustainable — Economic incentives align node operators, webmasters, validators, and users.
- Resilient — No single point of failure; the network grows stronger with each new node.
We are at Phase 2 — the testnet is live and public, nodes are running, and the core protocol is proven. Just as Mars needs thousands of colonists to build a new world, the internet needs thousands of node operators to build a search engine free from centralized control.