Pick a LumaUna project linked to GitHub, scan the repo if needed, then click a file to load it into Existing code (same workflow as opening a project in Cursor).
Select a project with a GitHub repo
Tip: Create projects under Projects if you haven’t yet.
Coding Agents
Loading coding model settings...
Agent
Strength
Benchmark
Use Case
Connected
Pipeline Steps
Generated Code
Proof
✎ CODE REVIEW RESULTS
Multi-model code review — issues found by 2+ models are marked high confidence.
▦ PORTFOLIO
All projects at a glance. Status, active tasks, pending approvals.
Project
Type
Tasks
Active
Pending Approval
Last Activity
Loading...
❖ INITIATIVES
Track cross-repo changes. See which repos are done, pending, or blocked.
📚 KNOWLEDGE
Global platform knowledge — shared across all projects. History, audit files, rules, and lessons learned.
■ PLATFORM KPIs
8 LumaUna-wide KPIs measuring verified delivery quality across all projects.
⚙ PROJECTS
Link projects to GitHub repos. Persistent memory tracks decisions, bugs, and progress across sessions.
Status
Loading...
GitHub
—
Files
—
Quick workflow① Settings → GitHub URL
→
② Scan Files
→
③ Tasks → New Task
→
④ Approve & Merge[hide]
Memory Summary
Click "Sync Memory" to load
Recent Commits
Loading...
AI Summary
AI proposed this reorganized memory:
Loading...
Status
Loading...
Run AI Audit
Select one or more AIs. Each will write its own audit file to .lumauna/ in GitHub and log to history.
Propose Cleanup
Consult up to two AIs to compare proposals for cleaning up memory, history, and rules. Nothing is saved automatically.
and optionally
Loading...
Audit Files
Each configured AI's audit file from .lumauna/. Click to expand.
0 files pendingstaged on server — not yet in GitHub
Click Run Tests to execute the configured test command in the staging directory.
Select a model and click Review to get an AI opinion before pushing.
Push to GitHub
ID
Description
Status
Priority
Models
Score
Actions
Task Log
Loading...
All generated files live here on the server. Edit directly, run commands, test — no GitHub needed.
When ready, push to GitHub from the Tasks tab.
Multi-model code review — issues found by 2+ models are marked high confidence.
See what changed since last deploy, check GitHub Actions status, then deploy when ready.
Last deploy
Unknown
These rules are injected into every AI chat and task. Edit to control AI behaviour.
Loading...
Project Knowledge
Project-specific entries. Global knowledge is in the Knowledge page.
Start chatting about your project...
GitHub connection
AI Audit Slots
Configure which AIs audit this project (up to 9). Each gets its own audit file in .lumauna/. All share history.md.
Work Mode
Immediate: every task commits directly to a GitHub branch (current default). Server-first: tasks write files locally — fast, no GitHub calls. You review (diff, tests, AI review) then push when ready. LumaUna Only: tasks build in an isolated workspace on this server. Nothing touches GitHub until you type "make it live".
▲ Make it Live
Push all approved workspace code to GitHub and/or deploy to the live server.
This is the only action that moves code out of LumaUna.
Services / Apps
Define the services that make up this project (e.g. frontend, backend, worker).
When queueing a task you can target a specific service — LumaUna will generate files in that directory and inject the right context.
£ USAGE & COSTS
API call tracking and cost breakdown by provider.
Connected Models
Provider
Model
Calls
Cost
Tokens In
Tokens Out
✉ MODEL API
All connected LLM providers, model versions, and endpoints.
⚙ SETTINGS
How each model is connected, plus 30-day usage by model.
API Keys
Keys are masked. Enter a new value to update server settings securely.
Service
Env Key
Status
Current (masked)
New Value
Action
Core Models (Operational)
Model
Provider
Strength
Weakness
Status
30d Calls
30d Cost
General Reasoning Catalog
Model
Provider
Strength
Weakness
Access
Connected
30d Calls
30d Cost
Latest Model Status (Live)
Checks provider APIs live and compares each configured area to the latest discoverable model.
Area
Scope
Provider
Selected
Live Latest
Status
Notes
Action
Mobile App Downloads
Latest Android and iOS app links. Refresh pulls from GitHub latest release and configured store URLs.
ANDROID
IOS
Platform
Channel
Version
Published
Source
Download
Voice Runtime Capabilities
Provider speech support for LumaUna: native realtime or server bridge fallback.
Provider
Voice
Realtime
Transport
Passkeys
Register a hardware or biometric passkey on this device for passwordless sign-in. Your browser or OS manages the key — nothing secret is sent to LumaUna.
Loading…
Chat Option Visibility
Choose what appears in chat selectors. Keep options concise for faster switching.
⧉ WORKFLOW LAB
Create multiple model workflows, run the same task across them, and audit which performs best.
Workflows
Loading...
Workflow Editor
Template Test Pack (6 samples)
Select a template to view sample problems and solution patterns.
Run & Compare
Run a workflow to see result
Audit Runs (Score 1-5)
When
Workflow
Task
Duration
Score
Action
LumaUna
Decision Platform
📚 LumaUna Support
Claude Sonnet · Ask anything about LumaUna
Hi! I'm LumaUna Support. I know everything about how LumaUna works.
Ask me anything — how tasks work, how to use server-first mode, what KPIs are for, how to set up audits, anything.