Changelog
All the latest features, improvements, and fixes to MemeScout.
Per-Wallet Entry MC Accuracy (Phase 2)
We now densify market-cap snapshots for tokens during their first 10 minutes of life — every minute, fresh tokens get an extra price+MC datapoint pulled from Birdeye. This gives the Smart Money entry-MC backfill enough granularity to distinguish a wallet that bought at second 5 from one that bought at second 200, instead of collapsing them onto the same t=0 snapshot. Leaderboard multipliers will reflect true per-wallet entry timing as new trades flow in.
Smart Money Leaderboard Filters
Quickly slice the Smart Money Leaderboard by review status. New filter chips let you view only Unreviewed, Checked, Retest, or Avoid wallets, plus dedicated filters for wallets you've added notes to or saved to your watchlist. Each chip shows a live count so you can see at a glance how much research is left to do.
Free Registration & Referral Rewards
Anyone can now create a free account to access basic features and participate in the referral program. Earn tiered rewards for referring paying subscribers: Bronze (3 referrals = 1 week Degen free), Silver (10 referrals = 1 month Degen free), and Gold (25 referrals = 1 month Chad free). Track progress, claim rewards, and view your referral stats from the Account → Referrals tab.
WebSocket v1.1 — Snapshot Replay on Reconnect
Every new WebSocket connection now automatically receives a 'snapshot' event containing all active radar signals (up to 50, fully enriched). This eliminates the delivery gap caused by edge function recycling — reconnecting clients instantly catch up on any signals missed during disconnects. Clients can also send {"event": "replay"} at any time to request a fresh snapshot on demand.
Enriched signal_update Payloads
The WebSocket signal_update event now sends the full enriched payload (name, symbol, market_cap, liquidity, risk_level, intelligence scores, etc.) — the same schema as signal_new. Previously it only sent 5 sparse fields (address, radar_score, trigger_reason, peak_mc, call_tier), forcing bots to make separate REST calls for metadata.
Re-appearing Tokens Now Fire signal_new
Tokens that return to the radar after being hidden (e.g., recovering from a dip) now correctly fire a signal_new event instead of being silently ignored. Previously, only database INSERTs triggered signal_new — UPDATEs that un-hid a token were missed entirely, causing a delivery gap between the daily digest and the WebSocket feed.
Call Tier Classification (PRIME / EARLY / STANDARD)
Every radar signal now carries a call_tier label based on market cap at first detection. PRIME (< $100K MC) = earliest alpha. EARLY ($100K–$300K) = still early. STANDARD (> $300K) = confirmed momentum. Visible across the dashboard, REST API, WebSocket feed, and Discord alerts.
Full Pipeline Deployment Sync
Redeployed ingest_radar, api, and ws edge functions to ensure scoring logic, call_tier classification, and API responses are fully synchronized with the latest codebase. No more logic desync between code and live environment.
Discord Alerts: Call Tier Badges
Discord alerts now show a colored call tier badge (🟢 PRIME / 🟡 EARLY / ⚪ STANDARD) so you can instantly gauge how early a signal is without opening the dashboard.
Scoring v3 — Noise Reduction & Momentum Signals
Major scoring engine overhaul based on 30-day performance analysis. Raised minimum radar score from 36 to 42 and minimum market cap from $25K to $50K — cutting ~60% of low-quality 'garbage' calls. Added two new momentum bonuses: Volume Velocity (+5 pts when 5m/1h volume ratio ≥ 0.25) and Buy Acceleration (+5 pts when buy/sell ratio ≥ 2.5).
Loosened Danger Hard-Block (700 → 900)
Data showed 31 of 59 big winners were flagged 'Danger' by RugCheck. The hard-block threshold has been raised from risk_score > 700 to > 900. Tokens in the 701–900 range now receive a -10 penalty instead of being discarded, allowing high-risk/high-reward plays to surface.
Phase 3 Scoring Baseline (Admin)
New admin dashboard card comparing pre-v3 and post-v3 call volume and quality metrics side-by-side, so we can track whether the scoring changes are moving the needle.
Wallet Review Status (Tri-State Annotations)
Leaderboard wallets can now be marked with four private, per-user review statuses: ✅ Checked (green), ❌ Avoid (red), ⚠️ Retest (yellow), or unmarked. Click the status icon to cycle through states. Notes and statuses are private — each user sees only their own annotations.
Configurable Leaderboard Min Trades
Admins can now set the minimum unique-token trade count required for a wallet to appear on the leaderboard. Configurable from the Admin → Settings tab without redeploying. Wallets below the threshold automatically lose their display name.
Wallet Watchlist & Custom Names
Save wallets from the leaderboard or manually add any Solana wallet address to your personal watchlist. Rename wallets to match your copy trading bot names. Manage everything from Account → Watchlist.
Private Wallet Notes
Add private notes to any leaderboard wallet. Notes are per-user and persist across sessions — only you can see yours.
DexScreener Trade Pattern Links
Each pick on the leaderboard now links to DexScreener with both the token and wallet address pre-filled, so you can inspect exact trade patterns directly.
Bot & Sniper Detection Overhaul
Tightened bot detection: wallets are now flagged if 80%+ of their trades have sub-3s buy delays, or if they show high-frequency patterns (e.g. 472 trades on one token). Flagged wallets are automatically hidden from the leaderboard and have their display names revoked.
Trade Deduplication & Transparency
Scoring now uses only the first trade per token per wallet, eliminating inflated trade counts from snipers. Each pick row shows the exact trade timestamp and buy delay (labeled 'before call' if ≤ 0s).
Show All Picks & 0% Win Rate Filter
Leaderboard now shows all picks per wallet with a 'Show all' toggle (instead of capping at 5). Wallets with 0% win rate are automatically hidden from the leaderboard.
Trade Count Pagination Fix (1000 Row Limit)
Fixed the scoring function silently capping at 1,000 trades due to Supabase's default row limit. All wallet trades are now paginated correctly, ensuring accurate stats even as the dataset grows.
Degen & Chad Tiers Now Available
Both the Degen and Chad subscription tiers are now open for purchase. Degen adds advanced filters, 15-second refresh, Quick Trade with Trojan, and Discord alerts (coming soon). Chad adds the Smart Money Leaderboard, API access (coming soon), and priority support.
Smart Money Leaderboard (Chad Exclusive)
The /leaderboard page is now gated to Chad tier subscribers. It displays anonymized 'alpha wallets' — wallets that consistently buy tokens within minutes of a radar call and achieve high multipliers. Each wallet gets a unique creative codename and color-coded avatar. Rankings are based on a composite Alpha Score weighing win rate, trade volume, and average multiplier.
Updated Tier Feature Alignment
Revised feature lists across all tiers for clarity: Ape focuses on dashboard essentials, Degen adds speed and trading tools with Discord alerts, Chad adds Smart Money intelligence and API access. Removed watchlists and custom alerts references. Help page tier descriptions updated to match.
Daily Digest Landing Page (/digest)
New dedicated newsletter signup page at /digest with a degen-styled layout showcasing the benefits of subscribing: daily alpha drops, rug alerts, win summaries, and weekly insights. Includes a live email preview mockup and stat highlights. Linked from the footer and all major pages.
Digest CTAs on /calls and /stats
Added 'Subscribe to the Digest' call-to-action banners on the Calls Archive and Stats pages to drive newsletter signups. The Stats page newsletter inline form was replaced with a link to the dedicated /digest page for cleaner tracking.
Stats Page Total Calls Count Fixed (1000 Cap Removed)
The Stats page was incorrectly capping Total Calls at 1000 due to Supabase's default row limit. Fixed by using an exact count query for the total and paginating through all records in 1000-row chunks (up to 10,000) to ensure win-rate percentages (2x, 5x, 10x) are calculated against the full historical dataset.
Minimum Radar Score Raised: 30 → 36
Data analysis of 81 calls on March 24th revealed that tokens scoring 30–35 accounted for 14 of 18 flatlined calls (tokens that never exceeded their entry price). Tokens with score ≥ 36 produced only 3 flatlines from 42 calls, pushing the 1.25x win rate from 59.3% to 78.6% on the same dataset. The minimum radar score has been permanently raised from 30 to 36 to reflect this finding and improve overall signal quality.
Score Distribution Analysis (3/24 Backtesting)
Full statistical review of 3/24 call data: 81 total calls, 18 flatlined (score 30–35 dominated), 25 hit 2x+. At score ≥ 36: win rate jumps to 78.6% (1.25x+) and 47.6% (2x+). The buy/sell ratio proved non-predictive for flatlines (0.577 vs 0.600) — score threshold was the strongest separator. This data will inform ongoing monthly threshold reviews.
RugCheck Score=1 False 'Good' Bug Fixed
Critical fix: RugCheck returns risk_score=0 or score=1 for tokens it hasn't indexed yet (i.e. very new tokens). Our system was incorrectly treating this as a legitimate 'Good' rating and awarding a +20 bonus score — meaning almost every new token was getting an unearned boost. Tokens with score ≤ 10 are now treated as unverified (no bonus, capped at 65), preventing inflated scores and bad calls on unindexed tokens.
Sell Pressure Hard Block
New radar filter: tokens where sells > buys × 2 (i.e. the sell ratio exceeds 66%) are now hard-blocked from entering the radar. This prevents calling tokens that are actively being dumped at discovery time.
Markdown for Agents (AI-Readable Pages)
Every page on MemeScout now has a machine-readable Markdown version for AI agents and LLM crawlers — inspired by Cloudflare's Markdown for Agents feature. Access any page as structured Markdown via the endpoint: /functions/v1/markdown-for-agents?path=/pricing (or /help, /changelog, /docs/api, /stats, /calls/:date, /token/:address). Dynamic pages query the live database so AI agents always get real-time data. Responses include YAML front matter (title, description, URL) and an x-markdown-tokens header for token estimation — up to 80% fewer tokens than raw HTML.
AI Crawler Discovery via Link Tags
Key pages now include <link rel="alternate" type="text/markdown"> tags in the document head so AI crawlers and agents can automatically discover the Markdown endpoint without manual configuration.
Updated robots.txt & llms.txt for Agent Discovery
robots.txt and llms.txt updated to advertise the new per-page Markdown endpoint, making MemeScout fully discoverable by LLM-powered tools, AI agents, and automated research pipelines.
On-Chain Rug Detection (No Indexer Delay)
MemeScout now fetches on-chain token metadata via Helius DAS at detection time, before GoPlus or RugCheck have indexed the token. Checks: Mint Authority (dev can inflate supply), Freeze Authority (dev can freeze your wallet), Metadata Mutability (dev can rename/rebrand mid-pump), and Top-3 Wallet Concentration (coordinated dump risk). Stackable penalties up to −50 radar score points.
On-Chain Signals on Token Detail Page
Token detail pages now show a dedicated 'On-Chain Signals' card sourced directly from the database (populated at ingest time, zero latency for users). Displays Mint Authority, Freeze Authority, Metadata Mutability, and Top-3 Wallet Concentration with clear Safe/Danger indicators.
Discord Alert Enrichment for Rug Signals
Discord alerts now include 🚨 MINT NOT REVOKED and 🧊 FREEZE ACTIVE warnings directly in the embed description. Alerts with active mint or freeze authority also use a red embed color to stand out from normal signals.
Parallel Helius Calls
Holder count fetching and the new on-chain getAsset call now run in parallel per token, keeping the 2-minute ingestion cycle latency unchanged despite the additional data.
Phase 2a: Full-Strength Freshness Penalty
Restored freshness penalties to intended values: ≤30m: −20, 31-60m: −15, 61-120m: −10, 121-180m: −5. Data showed fresh tokens (≤1hr) hit 5x at just 1.1% vs 13.6% for older tokens — the previous softened values were far too lenient.
Phase 2a: Increased LP Warning Penalty
LP warning penalties increased from −10/−15 to −15/−20. LP-warned tokens hit 5x at 6.9% vs 18.1% for clean tokens — an 11.2pp gap that the old penalty wasn't reflecting adequately.
Phase 2a: Widened Risk Level Modifier
Risk modifier range widened from 35pts (−20/+15) to 45pts (−25/+20). Good risk tokens hit 5x at 19.3% vs 8.4% for Danger — the old range allowed Danger tokens to average-score higher than Good tokens due to volume/liquidity accumulation.
Volume & Trade Count Data Pipeline Fix
Fixed a critical bug where volume_1h, volume_5m, buys, and sells were stored as null instead of 0 for tokens with no recent activity. JavaScript's || operator was treating 0 as falsy — switched to ?? (nullish coalescing) so genuine zero values are preserved.
First-Detection Volume Capture
Radar hits now store first_call_volume_1h, first_call_volume_5m, first_call_buys, first_call_sells, and first_call_liquidity at the moment of first detection. These are included in the CSV export for future scoring calibration analysis.
Softened Freshness Penalty
Halved the freshness penalty across all age brackets to avoid over-punishing high-quality early tokens. ≤30m: −10 (was −20), 31-60m: −7 (was −15), 61-120m: −4 (was −10), 121-180m: −2 (was −5). Tokens that pass multi-source discovery and risk filtering now score higher even when young.
Better Early Token Coverage
Combined with the Phase 1 scoring overhaul, the softened freshness penalty means fewer false negatives on promising early movers — cleaner signal-to-noise ratio overall.
Email Blast System
New admin tool for sending announcements to all registered users via email with per-user unsubscribe links.
Radar Score Phase 1 — Freshness Penalty
Replaced the freshness bonus (+8 to +15) with a tiered freshness penalty (0 to −20). Newer tokens are now penalized instead of rewarded — data showed fresh listings hit 5x at just 2.6% vs 18.7% for tokens that survived longer within the 4-hour window.
Radar Score Phase 1 — Risk Level Modifier
Risk level now directly impacts the radar score: 'Good' tokens receive +15 points, 'Warn' is neutral, and 'Danger' tokens receive −20. Previously, Good tokens only got +5 and Warn tokens were penalized −15 while Danger had no score impact.
Score Threshold Lowered for Validation
Minimum radar score temporarily lowered from 55 to 30 to capture more data during the Phase 1 validation period. This will be re-evaluated in March 2026 once enough data is collected under the new scoring formula.
Admin Data Export (CSV)
New Export tab in the admin dashboard for downloading enriched radar call data as CSV. Includes token info, performance metrics, risk indicators, intelligence scores, and market data — designed for offline correlation analysis.
Intelligence Scores (API + WebSocket)
New computed fields for smarter signal filtering: lp_health_score (0-100), volume_consistency (0-100), wallet_risk_score (0-100), and time_to_peak_estimate (fast/medium/slow). Available on /signals, /token/:address, and WebSocket signal_new events.
Digest-Worthy Filtering
Intelligence scores are computed during ingestion and stored in snapshots for historical analysis. Use these to filter to only high-quality, 'digest-worthy' tokens in real-time.
Real-Time WebSocket API
New WebSocket endpoint for instant radar signal delivery. Connect to receive signal_new events in real-time as tokens hit the radar — no polling required. Includes heartbeat keep-alive and connection management.
WebSocket API Documentation
Added comprehensive WebSocket documentation to the API docs page with connection examples in JavaScript and Python, event reference, and connection limits.
Signal Status Indicators in API
Added is_new, is_returning, is_recovered, and is_dropped boolean flags to the /signals endpoint, allowing API consumers to replicate dashboard status labels without additional client-side logic.
Renewal Reminder Email Delivery
Fixed subscription renewal reminder emails failing to send due to incorrect sender domain (solanaradar.xyz → memescout.io).
Historical Calls API Endpoint
New GET /calls endpoint for accessing historical radar call data. Query by date or last N days, filter by minimum multiplier, and analyze call performance programmatically.
Trading Signals API
REST API for Degen and Chad tiers to access radar signals programmatically. Generate unique API keys, rate-limited access, and interactive documentation at /docs/api.
API Key Management
New API tab in Account settings to generate, view, and revoke API keys. Keys are SHA-256 hashed for security.
Interactive API Documentation
Comprehensive API docs with code examples in Python, JavaScript, and cURL. Includes endpoint reference and error handling guide.
Tiered Market Cap Discovery
Age-based MC limits: $5M max for tokens under 1 hour, $2M for 1-4 hours, $800K for older tokens. Captures hot launches previously filtered out.
Bubble Maps Integration
Added holder distribution visualization via Bubblemaps on token detail pages with a tabbed interface alongside the price chart.
Bubble Maps Display Fix
Fixed holder map not loading for older tokens by switching to the official Bubblemaps iframe endpoint.
Enhanced Discord Alert Links
Added Solscan blockchain explorer link to Discord alert buttons for direct on-chain verification.
Archive Token Search
Search tokens by name, symbol, or address across all historical calls. Results show entry MC, peak MC, multiplier, and call date.
Discord Alert Age Display
Fixed token age showing '56 years ago' in Discord alerts due to incorrect timestamp calculation.
Mobile-Friendly Contract Address
Changed contract address formatting in Discord alerts from code blocks to inline code for easier copying on mobile.
Discord Webhook Alerts
Automated alerts sent to Discord for new radar calls, featuring token details, metrics, risk levels, and quick-links to DexScreener, Birdeye, and RugCheck.
LP Provider Warning System
Visual warning badges for tokens with limited liquidity providers - shows amber indicator with score penalties instead of blocking tokens entirely.
Human-like Tweet Posting
Reduced automated tweet frequency to 12/day maximum with randomized delays to prevent account lockouts and appear more natural.
More Tokens on Radar
LP-related risks no longer block tokens from appearing - they now show as warnings with score penalties, increasing radar discovery volume.
Tour Completion Celebration
Added confetti animation and Scout mascot celebration when completing the dashboard tour.
Aggressive Filter Banner Tour Step
New tour step explaining the high-risk filter warning banner for better user guidance.
Enhanced Tour Experience
Scout mascot now appears in the tour prompt modal for a friendlier onboarding experience.
Dashboard Tour Restart
Users can now restart the dashboard tour anytime from the help section or tour prompt.
Tour Prompt Modal
New modal prompts first-time users to take the interactive dashboard tour.
Tour Step Improvements
Added more detailed explanations for each dashboard feature during the tour.
Interactive Dashboard Tour
Step-by-step guided tour of the dashboard using Driver.js for seamless onboarding.
Scout Mascot Integration
Introduced Scout, the friendly meerkat mascot, throughout the app for personality and guidance.
Empty State Improvements
Scout now appears when no tokens match filters, providing helpful suggestions.
Notification Sound Controls
Added browser notification sounds for new token alerts with customizable volume.
Filter Presets System
Save and load custom filter configurations for quick access to preferred settings.
Hotness Meter Visualization
Enhanced visual feedback for token momentum with animated hotness indicators.
AI Token Analysis
Integrated AI-powered token analysis providing insights on tokenomics and social signals.
Rug Check Safety Scoring
Added safety scores from RugCheck API to help identify potential risks.
Token Age Display
Fixed incorrect token age calculations for recently launched tokens.
Token Holder Distribution
Added holder distribution charts showing top wallet concentrations.
Aggressive Filter Warning
Visual banner warning when high-risk filters are enabled to prevent missed opportunities.
Dashboard Load Optimization
Reduced initial dashboard load time by 40% through lazy loading.
Calls Archive
Browse historical token calls organized by day with performance tracking.
Peak Market Cap Tracking
Track the highest market cap reached by each token since radar detection.
Mobile Responsive Tables
Improved token table layout for better mobile viewing experience.
Solana Wallet Payments
Pay for subscriptions directly with SOL using Phantom, Solflare, or other Solana wallets.
Referral Program
Earn rewards by referring friends with unique referral codes and tracking.
Authentication Flow
Fixed edge cases in magic link authentication and password reset.
Token Detail Pages
Comprehensive token pages with charts, stats, holder info, and external links.
Stats Dashboard
Performance statistics showing call accuracy and multiplier tracking.
Real-time Updates
Live token data updates without page refresh for active monitoring.
Initial Dashboard Launch
Core token radar dashboard with filtering, sorting, and real-time data.
Token Filtering System
Filter tokens by market cap, liquidity, volume, age, and radar score.
User Authentication
Secure signup and login with email/password and magic link options.
Subscription Tiers
Free, Basic, Pro, and Enterprise subscription plans with feature gating.