Syncing…

Feature Requests 0

Loading…

Custom Agents
Specialist agents that join the delivery pipeline for features they're assigned to.

AI Squads 0

🤖

No squads yet.

Configuration

API keys and integration settings. Secrets are never sent back to the browser.

Settings Source
All configuration is stored in the database. Changes take effect immediately without restarting Nova.
☁️ Database
🤖 AI Providers ▶
ANTHROPIC_API_KEY —
OLLAMA_URL —
🗄️ Database & Memory ▶
DATABASE_URL —
MEMORY_API_KEY —
DASHBOARD_PORT —
💬 Slack PO Bots ▶

Each squad PO needs its own Slack app. Each app has a unique @mention alias.

Sets SLACK_PO_LLM. Claude requires a valid API key. All others require Ollama running locally.
Changes here are reflected in SLACK_PO_INSTANCES when saved to the database.
🔗 Atlassian (Confluence & Jira) ▶
ATLASSIAN_BASE_URL —
ATLASSIAN_EMAIL —
ATLASSIAN_API_TOKEN —
☁️ Control Plane ▶
CONTROL_PLANE_URL —
CONTROL_PLANE_API_KEY —

Database Connection

Required for Upload / Load / Database source mode. Uses DATABASE_URL from above if left blank.

🪙 Alpha Credits ▶

Your token budget for this deployment. Credits are managed by your admin.

Balance
—
Used
—
Total
—

Concurrency Control

Configure concurrent job slots and rate limits. Changes apply immediately to the running process.

Live Pipeline (auto-refreshes via real-time events)
Session Token Usage
Claude In
—
Claude Out
—
Ollama In
—
Ollama Out
—
Est. Claude Cost
—
⚙ Queue Concurrency (concurrent job slots per stage) ▶
PO Planning —
BA Analysis —
Development —
QA & Testing —
Specialist —
Custom Agents —
🚦 Rate Limits & Job Config ▶
Claude Max Concurrent API Calls —
Ollama Max Concurrent Calls —
Job Timeout (seconds) —
Max Job Retries on Failure —

Notifications

Choose which events trigger outbound notifications. Configure Slack or webhook targets per event.

🔌 Channel Configuration ▶

Set a default Slack Incoming Webhook URL and/or a generic webhook endpoint. These are used when you enable Slack or Webhook for any event below.

Default Slack Webhook URL
Default Webhook URL
Webhook Secret (X-Webhook-Secret header)
Loading events…
Message Queue Log
No log entries yet.

Users

Create and manage Nova user accounts.

Username Role Squad Credits Actions
Loading…

Agent Permissions

Control how each agent is allowed to run. Changes are saved to the database and take effect within 30 seconds.

Execution Mode: supervised — runs normally  ·  allowed — runs with no restrictions  ·  dry_run — agent is skipped entirely (pipeline stops)
Loading…

Manage

User invitations and access management.

🔗

Invite to Nova

Generate a one-time link (valid for 1 hour) or send a direct email invitation.

Opens your default mail app with a pre-filled invitation. A new link is generated automatically if none exists yet.

Nova Intelligence

Nova Pool Analytics for Company A

Loading memory statistics…

CortexOS Architecture

Visual blueprints of the system's core design — memory, solution, communication, and LLM strategy.

🧠

Memory Architecture

Two-tier persistent memory: PostgreSQL + pgvector as primary store with JSON fallback. Embeddings generated via Ollama's nomic-embed-text model (768 dimensions).

🏗️

Solution Architecture

End-to-end pipeline from file drop or API to a delivered feature — watcher, queue, orchestrator, gated agent stages, and dual-write persistence.

📡

Communication Architecture

Queue-based async messaging with a 2-second worker poll loop. Human decision gates block stage progression until resolved via the dashboard.

🤖

LLM Strategy

Unified AI client routes per-agent LLM calls to either Claude (Anthropic cloud) or Ollama (local). Each agent's provider is configurable per feature.