Methodology
4-Layer Signal Architecture
SOLIS detects emerging Solana narratives by triangulating signals across four independent data layers, each with different lead times relative to market price action. For a full visual deep-dive, see the Brain Architecture page.
Layer 0: Social (LunarCrush + X/Twitter, opt-in)
Social sentiment and engagement signals. When community buzz precedes developer activity, it can signal early narrative formation.
- - Social interactions and engagement volume (LunarCrush)
- - Sentiment scores (bullish/bearish ratio)
- - Galaxy score and social dominance
- - X/Twitter KOL timeline signals (Mert, Toly, Akshay, Messari, a16z, Solana Devs)
- - Topic extraction via cashtag and protocol name matching
Layer 1: Leading (GitHub)
Developer activity that precedes market movement by 2-4 weeks. When devs start building, the market hasn't noticed yet.
- - Star velocity on 64 curated Solana repos + dynamic discovery
- - Commit frequency surges (z-score > 2.0)
- - New repo clusters by topic
- - Contributor count deltas
- - Fork rate anomalies
Layer 2: Coincident (DeFi Llama + Helius)
Real-time capital and onchain activity. When TVL and transaction volume align with dev signals, the narrative is materializing.
- - Solana chain TVL history and protocol-level changes
- - DEX volumes and fees
- - Stablecoin flows (net inflows/outflows)
- - Program activity (transaction volume per program)
Layer 3: Confirming (CoinGecko)
Market price and volume data. When tokens in a narrative cluster start moving, the signal is confirmed but alpha is diminishing.
- - Solana ecosystem token prices and volumes
- - Category market cap movements
- - Trending coins
Signal Stage Classification
| Stage | Layers Active | Alpha Potential |
|---|---|---|
| EARLY | Layer 1 only | Highest — devs building before market notices |
| EMERGING | Layer 1 + 2 | High — builders and capital moving together |
| GROWING | All 3 layers | Moderate — gaining mainstream traction |
| MAINSTREAM | All 3, high confidence | Low — likely already priced in |
Anomaly Detection
SOLIS uses z-score analysis to identify statistically significant deviations from baseline activity. A z-score above 2.0 (2 standard deviations) flags a metric as anomalous. No machine learning is involved — it's pure statistical math applied to time-series data.
LLM Analysis
SOLIS uses exactly 2 LLM calls per report via pure TypeScript orchestration (no Agent SDK):
- Claude Haiku 4.5 (via OpenRouter) — Primary model for narrative clustering and build idea generation. Temperature 0.3 with JSON schema enforcement.
- GLM-4.7 → GLM-4.7-flash (fallback) — Fallback chain activated on 5xx errors. Cheaper but less capable.
Total LLM cost per report: ~$0.06. At daily cadence, ~$22/year.
Data Sources
| Source | Access | Rate Limit |
|---|---|---|
| GitHub API | Free (PAT) | 5,000 req/hr |
| DeFi Llama | Free (no key) | No hard limit |
| Helius | Free (1M credits/mo) | ~1M RPC calls/mo |
| CoinGecko | Free (10K/mo) | ~330 calls/day |
| LunarCrush | Free tier (opt-in) | Configurable throttle |
| X/Twitter | Pay-as-you-go (opt-in) — KOL timeline tracking | Per-user rate limits |