Drag tickets to change status — click any field to edit inline. No save button.
Project
Board
List
Prioritize
Workflow
Chats 0
Prompts 0
Project Settings
Synthetic Data
Workspace
Investments 3
Businesses 2
Projects 61
Tool Settings
Settings
API Keys
Delivery Tiers
Agents
AI Engines
Workflow
Webhooks
Navigation
Detect Tool
0
Prioritization — drag to rank
Top = highest priority. Agents read rank to decide what to pick up next. Reordering updates rank instantly.
Delivery horizonParallel agents10 items / day • Day delivery horizon1 agent • 10/day — add agents for 1.7× diminishing returns (trip-over).
Tip: this is backlog ranking. Higher-ranked tickets should be worked first. MCP exposes rank via search.
61
Directory: click a card/row to switch project. Star is coming soon. Right-click a project in the dropdown to edit.
0 prompts
Add prompt to NORA
Linked to project • will be searchable via prompts ~ text
📋 Paste entire historical chat dialog
Paste a full transcript (e.g. Cursor/Replit chat export). AI will extract each requirement-oriented prompt and nest revisions under their main feature request. Uses your Meta Muse API key (localStorage nora:metaMuseKey) or heuristic fallback.
Hints → Meta Muse • Bugs & Sample Records labelled • “start webserver” hidden
↗ Chat link
Fetch follows the link server-side (avoids CORS) and extracts the full chat dialog into the paste box above — same flow as copy/paste for Extract.
Decompose creates an Epic/Feature/Story/Task from a prompt and links them (is a defect of/has defect etc. not needed — prompt ↔ ticket link is stored on the prompt).
Link attaches an existing ticket. Use JQL prompts ~ "sniper" via MCP to find prompts.
0 chats
📼 Archive full chat — rich text, versioned
Paste with formatting intact (colors, bold, code blocks). The raw HTML is stored versioned per chat, preserving tool-identifying chrome. Text is also extracted for search/AI. Attachments/media placeholder below — coming soon. This sits alongside Prompts (parsed, typed requirements).
Rich HTML archived • plain text extracted for Extract & JQL chats ~ text
📎 Attachments/mediacoming soon— placeholder. Future: images, files, audio linked to this chat version. For now, paste images inline in the rich box above (they will be stored as data URLs).
Versioned: each save creates a new version (v1, v2…) — history is kept. Extract on a chat parses its plain text into Prompts (typed Requirement/Bug/Sample). Chats are searchable via chats ~ "text".
Help & Guides
Ticket types, settings, and how agents & humans integrate via MCP or REST.
⌘ secret splitPrompts ↔ List/Board — derived lineage • drag header to swap • ⌥/⌘-click List/Board/Prompts again to exitView:
Aligned grid — left join: each left row aligned to its prompt(s)/artifact(s) on right. Swap flips left/right.
Nested — prompts nested under stories or stories nested under prompts (toggle via Swap). Click parent to expand.
Tip: ⌥ Option (or ⌘) + click List/Board then Prompts (or vice-versa) decides left/right. Artifacts shown as List (type, not status). Use Swap or drag headers to flip. Grid = left join aligned, Nested = hierarchical. derived links show prompt IDs only.
STORY
NORA-1
Create ticket
Accepted and expected for Business Process tickets. Any Mermaid diagram is supported — live preview renders when you view the ticket. See mermaid.js.org.
After creation, click any field in the ticket view to edit inline — no save button, just click, type, and blur.
Project
Format: Name | Description per line. Example POS: Store Associate, Store Manager, IT Admin, Finance Controller, Store Operations Team Member. Used when AI generates stories/features/epics from prompts (Gherkin).
?
Generates 16×16 pixel art via Meta Muse API (key in localStorage nora:metaMuseKey) or deterministic fallback. Upload overrides.
Workflow builder
✦
AI Live — requests & responsesMessenger bubbles: blue → request • green ← response • amber = pending
no calls0 completed
Live inspectorPending at top, newest completed next — request (right, blue) / response (left, green) • colors make it scannable like messenger
LLM Debug — Meta Muse requests & responses
What you’ll see here
Every call to Meta Muse (prompt extraction + story generation + tool detection + icon) is logged with full request JSON (what nora sent) and raw response (what Muse returned), plus endpoint, engine, key presence (masked), HTTP status and duration. If generating stories shows heuristic fallback, check here — it means the LLM call failed (no key, wrong key, or network/CORS) and the log will show the error.
Model in use: muse-spark-1.2-contributor via https://api.meta.ai/v1/chat/completions (or /api/meta-muse-chat proxy when running via node server.js to avoid CORS). Tool Settings controls per-feature engine; default is Meta Muse.