SupplyCore Console

Settings

Business settings

Keep settings focused on business choices, data freshness, and only show deep runtime detail when you need it.

Live refresh · Never

Off

Market anomalies

3 active alerts

View all

Critical misprice Alliance Market

Plutonium Charge S

Current 0.02 ISK · Normal 833.01 ISK · 0.00% of normal

1035466617946 · Fresh 3m ago

Critical misprice Reference Hub

Nanofiber Internal Structure I

Current 4,656.00 ISK · Normal 186,681,434.85 ISK · 0.00% of normal

60003760 · Fresh 1d ago

Critical misprice Reference Hub

Microwave M

Current 5,900.00 ISK · Normal 14,684,450.53 ISK · 0.00% of normal

60003760 · Fresh 3m ago

AI & Alerts

AI briefing behavior, report controls, deal alert thresholds, and killmail intelligence inputs.

AI Briefings Report Management Deal Alerts Killmail Intelligence

Configured Mode

Enabled

Provider: Groq (Free tier).

Connection Status

Ollama reachable

Status is checked against the configured local Ollama API endpoint.

Scheduler Behavior

Non-blocking

Disabled or unreachable AI falls back to deterministic summaries instead of blocking the job.

Capability Tier

MEDIUM

manual override: medium

Candidate batching, prompt depth, enabled tasks, and dashboard richness all scale from this centralized AI strategy.

Enable AI doctrine briefings Turn this off to force deterministic fallback summaries while still keeping briefing records populated.

AI Provider

Use Local Ollama for self-hosted CPU/GPU inference, Runpod Serverless for async GPU jobs, or Claude API for fast hosted inference (requires separate API billing at console.anthropic.com).

Local Ollama API URL

Use the API base URL, for example http://localhost:11434/api.

Runpod Serverless Endpoint

Paste the full Runpod async request URL, for example https://api.runpod.ai/v2/58qz2qbho8h3f1/run. Existing /runsync URLs are converted automatically.

Runpod API Key

Saved key preview: rpa_••••••••••••••••••••••••••••••••••••••••••nwkp.

Claude API Key

Get your API key from console.anthropic.com. API usage is billed separately from Claude Pro subscriptions.

Claude Model

Recommended: claude-sonnet-4-20250514 (fast, cheap) or claude-haiku-4-5-20251001 (fastest, cheapest).

Groq API Key

Saved key preview: gsk_••••••••••••••••••••••••••••••••••••••••••••••••B87k.

Groq Model

Recommended: meta-llama/llama-4-scout-17b-16e-instruct (500K TPD) or llama-3.3-70b-versatile (best quality, 100K TPD).

Model Name (Ollama/Runpod) Request Timeout (seconds) Capability Tier

Leave this on auto to infer capability from the configured model name, or pin a tier if the model naming does not include parameter size.

Theater AAR Prompt

Custom prompt for generating Theater After Action Reports. The battle data JSON (alliances, fleet composition, notable kills, top performers, turning points, ISK stats) is automatically appended. Leave empty to use the built-in structured AAR template with 8 sections (Executive Summary, Battle Overview, Fleet Composition, Key Turning Points, Tactical Assessment, Performance Insights, Risk Signals, Recommendations).

Configure either the local Ollama API or the Runpod serverless endpoint here, then manage cadence under Settings → Automation & Sync for the rebuild_ai_briefings scheduler job. Runpod requests now submit asynchronously and poll for completion within the configured timeout window. Small tiers stay compact, medium tiers add explanation and deltas, and large tiers unlock richer operator briefings while still keeping deterministic calculations authoritative.