Dev Console is GNI's technical transparency hub, designed for developers, system architects, and anyone who wants to understand how GNI works under the hood. Every component of the autonomous pipeline is observable here — from the 66-pattern injection security filter to the A/B prompt quality system and the self-healing adaptive engine. The API reference documents all 21 endpoints with method, path, and description, making it straightforward to integrate GNI intelligence into external systems. Source health monitoring tracks all 25 RSS feeds in real time, showing which sources are healthy, degraded, or down and how their trust weights are adjusted dynamically. The quota monitor ensures GNI never exceeds its free-tier token budget, with a safe ceiling at 85K tokens and a hard limit at 100K per day — the foundation of the $0.00/month architecture.
Full AI funnel trace showing how each article was selected stage by stage. View the complete pipeline from 150 raw articles down to the final 3 selected for analysis. Essential for understanding GNI’s editorial decision-making process.
The frequency controller, A/B prompt quality system, and L3.5 self-healing design that give GNI its autonomous capabilities. Shows how GNI adjusts its own behavior based on escalation levels and quota constraints. This is the architecture behind GNI’s L7 autonomy rating.
Real-time pipeline status with quality scores per run across 5 dimensions: relevance, depth, actionability, source diversity, and novelty. Includes A/B prompt test current state and historical quality trend. Use this to verify GNI is operating within expected quality parameters.
66 injection patterns actively blocked at Stage 1b of the pipeline with SHA-256 audit chain for tamper detection. Full pentest results and security audit trail available. GNI’s security layer ensures all 25 RSS sources are validated before any content reaches the AI analysis stage.
Live status of all 25 RSS sources: healthy, degraded, or down. Dynamic trust weights adjusted by GPVS prediction accuracy are shown per source. Mini sparkline charts reveal each source’s reliability trend over the last 7 days.
The self-healing log showing every time GNI’s heartbeat triggered the adaptive pipeline — including the reason (escalation delta), tokens consumed, and analysis result. This is GNI’s L3.5 self-healing evidence, demonstrating autonomous response to geopolitical threat changes.
Live Groq token budget showing usage per pipeline type vs the 85K safe ceiling vs the 100K hard limit. Historical usage trends and per-pipeline breakdown available. This page is the proof behind GNI’s $0.00/month architecture claim for the IEEE paper.
Three REST endpoints allow programmatic access to GNI's complete dataset: reports, predictions, and pipeline articles — all exportable as CSV or JSON. No authentication required — all data is publicly accessible. Essential for IEEE paper replication, external analysis, and building applications on top of GNI intelligence.
Autonomous web-layer health monitoring that runs every 30 minutes. Checks Supabase connection, report freshness, Groq quota, source health, and pipeline recency. Sends Telegram alerts automatically when issues are detected and saves all results to Supabase for historical analysis.