Aurea Ops
Release Execution Plan
Project: · Updated: · data · milestones doc · roadmap doc
exit criteria ↗
| Feature | Effort | Status | Owner | Notes |
|---|---|---|---|---|
| No features match the current filter. | ||||
Loading release execution state…
Loading spend data...
System Health
Playtest Runs adventure-quality harness · newest first · 🔴<3 🟡3–3.99 🟢≥4
node test-harness/quality-playtest/run.js --label=core --matrix=core| When | Label | Cells | Avg | Problems | $/turn | Total | Fails | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
||||||||||||
Spend Anomaly Detector (3σ z-score per purpose · 7d baseline vs 24h candidate · cron-anomaly-check)
No anomaly-detector history yet. Run
node proxy/audit-spend-anomalies.js or wait for the
next 6-hour cron firing.
| Purpose | Last verdict | z-score | Candidate avg | Baseline mean | Detected |
|---|---|---|---|---|---|
Flagged anomalies (last 14d)
✓ No anomalies in the last 14 days
Stuck-Session Detector (patterns A/B/C/D · 6h stale threshold · cron-daily)
No stuck-session scan data yet. Wait for the next cron-daily run (06:00 UTC),
or run node proxy/audit-stuck-sessions.js --write.
Affected adventures
Repair:node proxy/unstick-character.js <uid> <adventureId> --apply
| Pattern | UID | Adventure | Label | Last played |
|---|---|---|---|---|
✓ No stuck sessions detected
Adventure Quality (last 24h · cron-quality-rollup · dual-source) ⓘ
Alerts
✓ No alerts — all systems healthy
World Coverage
| World | Origins | Adventures | Thresholds |
|---|---|---|---|
Report History
gate2-bot-*, auto-play-bot*, playtest-bot-*. See proxy/api/v1/_test-bot.ts.
Daily Spend (Last 30 Days)
Daily Spend by Subsystem (Last 30 Days)
By Service
| Service | Calls | Cost | Avg/Call |
|---|---|---|---|
By Service / Model
| Service / Model | Calls | Cost | Avg/Call | % of total |
|---|---|---|---|---|
By Endpoint
| Endpoint | Calls | Cost |
|---|---|---|
Anthropic Billing (This Month)
Loading Anthropic cost data...
| Description | Cost |
|---|---|
Replicate Billing (This Month)
Loading Replicate cost data...
| Model | Runs | Succeeded | Cost |
|---|---|---|---|
Manual Billing Entries
For services without billing APIs — log actual spend from provider dashboards
| Date | Service | Amount | Tokens | Note | |
|---|---|---|---|---|---|
No manual entries yet
Player Leaderboard
Loading player data...
| # | Player | Day | Week | Month | Total | Cost | Adventures | Active | Done | Avg Turn | Mode | Platform | Last Active |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
No player data found.
Loading players...
No players match search
Select a player to view details
Loading player data...
Characters
No characters
Adventures
| Title | Character | Status | Turns | Style | Last Played | |
|---|---|---|---|---|---|---|
No adventures
Blueprint
Session
Turns
No turns yet
Quick Actions
Feedback Flags
No flags
Generate Image
Generate images with Seedream ($0.04) or GPT Image ($0.20). Choose model, size, and storage folder.
Session History
Origamify
Transform images into Faeria's origami paper-craft style via gpt-image-1.5 (~$0.20/image). Upload multiple files or paste URLs to batch process.
Storage Gallery
Browse images in Firebase Storage by folder. Click to view, copy URL, or delete.
Loading...
No files in this folder.
Workbench
Saved generated assets — a scratch space for images before they're assigned to stories
Loading...
No assets yet. Generated images auto-save here.
Loading stories...
· ·
No images yet
· ·
No images yet
No stories in this world yet.
No stories found.
Loading storage...
No images in storage yet.
Model Feed
Loading model feed from Firestore...
No models loaded. Click Refresh from sources to scan Hugging Face trending, or add models manually below.
Add Model Manually
Current:
Criteria:
| Model | Provider | Est. Cost | Status | Quality | Speed | Notes | ||
|---|---|---|---|---|---|---|---|---|
No models in queue. Star models in the Feed tab and click "Bakeoff" to add them here.
Comparison
| Prompt | |
|---|---|
|
/5
View output |
|
| Average | /5 |
Golden test sets for evaluating model quality. Each set contains reference prompts with expected characteristics and evaluation criteria.
Add custom prompt
No golden sets loaded. Click "Reload from Server" or import from a playtest JSON file.
Scene Setup
Configure a test scene, select writing styles, and generate side-by-side prose comparisons. (Writing style = the storytelling prose voice — Grounded, Earthsea, Heroic, etc. NOT the TTS engine.)
generating...
History
TTS Engine Bakeoff — side-by-side audio review
Compare TTS engines (Inworld, VibeVoice, OmniVoice, Kokoro, csm-1b) by playing the same fixture text rendered through each. Rate 1-5 to persist your scores; aggregates show overall winners. Renders are produced offline by proxy/data/voice-bakeoff/run-bakeoff.js — this tab is the listening surface.
No bakeoff runs found yet.
Run cd proxy && node data/voice-bakeoff/run-bakeoff.js to generate one.
Lines of Code
Growth over time by language. Data from loc.sh, backfilled from git history.
Cross-project totals overlay. Each line = one project from /ops/loc-projects.json. Add a project: append a registry entry.
| Project | Status | Snapshots | Latest | Total |
|---|---|---|---|---|
| OK — |
| Date | Build | Dart | TypeScript | JavaScript | Shell | Total | Files |
|---|---|---|---|---|---|---|---|
Loading config...
Identity & Preamble
Rules
Individual rule lines injected into the narrator prompt.
Sections
Multi-line rule blocks (combat, pacing, prose craft, etc.)
Childhood Guardrails
Tavern Chat Moderation
Queue over chat_moderation_review — filtered messages + invitations.
Items:
unactioned
Loading moderation queue…
Queue is empty — nothing to review.