Nova AI replaces Monday.com with a tool the team shapes together — then adds an AI layer that reads your tickets, writes the fix, and waits for your approval before merging.
Replace Monday.com. Give the team a tool they use daily and help shape. Issues, sprints, collaborative planning, and a personal space — all backed by GitLab, nothing stored in Nova.
All open issues across every game, grouped by workflow status. Filter by game, priority, type, assignee. Toggle between a list view and a swimlane board with one row per game × one column per status.
Walk through →Sprint health across all games at a glance. Each game row shows a burndown sparkline, health badge, and issue counts. Blocked games surface first. Expand any row to see its tickets grouped by status.
Walk through →Structured collaborative planning for anything — new game proposals, engine upgrades, feature specs, or sprint goals. Multiple people contribute, AI helps identify gaps and risks, required approvers sign off. The result is a detailed plan or spec that can be handed to developers, split into tasks, or passed to AI for execution.
Walk through →Real-time triage queue. Issue assignments, MR review requests, planning approvals, and stale doc flags arrive here. Add to Today, schedule, or dismiss. MR reminders escalate from gentle nudge to team lead CC.
Walk through →Everything you need to do today in one place. AI suggests items each morning — stale MRs, pending approvals, overdue issues. You choose what to add. Planned view shows your calendar; backlog is your parking lot.
Walk through →Escalating reminders for unreviewed merge requests. Gentle at 24h, direct at 48h, adds ticket owner at 72h, team lead CC at 96h+. Snooze with a reason. Respects business hours.
Walk through →Invite-only user management with four roles (Owner, Admin, Member, Viewer). Integration connections for GitLab and Drive. Workspace settings for timezone, business hours, and defaults. In Phase 2: AI model configuration per task and token budget controls.
The core promise. Describe a ticket, AI reads the codebase and writes the fix, opens a merge request, and waits for your approval. The AI never merges — that gate is structural, not a policy.
Click Execute on any ticket. Claude reads the codebase, generates a plan for your review, implements the approved plan, opens a GitLab MR, and waits. You review, request revisions (AI fixes them), and merge when ready. Includes AI code review, execution persistence with crash recovery, and a token budget guard.
Walk through AI executing →Describe an issue in plain English — with screenshots if you want. AI structures it into a proper ticket with title, type, priority, labels. Then triage intelligence suggests assignees based on code ownership and flags potential duplicates.
Walk through →AI reads the MR diff and generates a structured test brief: areas affected, prioritised test cases with steps, edge cases, regression risks. Dev approves, QA tests against a locked commit. New commits auto-invalidate the brief.
Walk through →Gantt-style delivery timeline with certification pipeline stages. Full game lifecycle from Dev → Internal QA → Spark → GLI → US state certs → Release. Inbound test completion emails auto-parsed and mapped to pipeline stages. Quarter navigation, drag-to-plan, stage-coloured bars.
Walk through →Deploy tracking across GCP and Spark targets. Automated Spark release emails. Test cycle tracking. Framework dependency matrix showing which games are behind on engine versions.
Walk through →Cross-project view of urgent issues grouped by game, sorted by severity. Blocked games at top. The "what's on fire" view that replaces the morning standup panic.
Walk through →The full platform. Per-game data hubs, Gemini-powered docs, smart features, and power-user tools. Everything connected — cheats, translations, sounds, game specs, release history — all in one place per game.
Every game gets its own hub: live micro-client, game spec from the math GDD, cheats editor with 46 force tools across 8 categories, side-by-side translation editor for 16 locales, sound brief management with waveform playback and revision threads. All real data.
Walk through game hub →Confluence-style knowledge base backed by GitLab wikis. Gemini answers questions across everything with section citations. Nightly stale detection flags docs that have drifted from the code.
Walk through →Workspace-wide metrics: total issues, completion rate, blockers, project health. Company news feed with team announcements. The home screen when you open Nova.
Walk through →This month: $31 of $150 budget — 21%. Nova costs scale with output (AI executions), not headcount. No per-seat fees. Compare: $31/month vs $200+ for per-seat PM tools with a team of 10.
Walk through →Press ⌘K from any screen. Fuzzy search across issues, projects, docs, and AI actions. Keyboard chords skip menus: C→T creates a ticket, G→R jumps to roadmap. Power users never touch the mouse.
Walk through →Type / in any text field to open a contextual command menu. Formatting, templates, inline linking, AI commands, and metadata shortcuts — all from the keyboard.
Searchable snippet library stored in a GitLab repo. Folder tree, syntax highlighting, Gemini semantic search. Reusable code patterns the whole team can find and share.
Walk through →Link Miro boards to tickets and planning threads. AI reads board content as context when executing. Design mockups, architecture diagrams, and brainstorms — all connected to the work they describe.
Team management features. Absence tracking, reviewer pool intelligence, engine release coordination, and sound asset management across all game projects.
Team availability calendar. When someone's away, their reviews auto-reassign, their planning thread approvals route to delegates, and sprint capacity adjusts automatically.
Smart reviewer assignment based on code ownership, current workload, availability, and expertise. Ensures reviews are evenly distributed and matched to the right people.
Audio brief tracking per game. Producers create briefs, audio engineers deliver to Drive, devs integrate. Status, versions, and conversation threads in one place per sound asset.
Walk through →When the engine team cuts a release, all frontend devs on dependent projects are notified with release notes, API changes, and affected files. Roles defined per-project in configuration.
Nova builds itself. The self-deploy pipeline closes the loop — approved code deploys automatically. Multi-agent architecture enables specialised agents to handle different types of work.
MR merged → Cloud Build triggers → deployed to Cloud Run. Nova ships its own features through the same execution loop it provides to users. New tools, integrations, and workflows are added by logging tickets — the platform extends itself.
Infrastructure for specialised AI agents — each with its own instruction set and tool permissions. The Dev agent (code execution) ships first and is already proven. Additional agents for planning, QA, and other workflows plug into the same framework as capabilities mature.
The AI can't merge, delete, or deploy. These capabilities don't exist in its tool set — safety comes from absent tools, not behavioural promises.
Code stays in GitLab. Docs stay in Drive. Nova stores only metadata — execution state, notifications, preferences. Your content never lives in our database.
No per-seat pricing. AI costs are usage-based. Runs on the team's own GCP. One billing account, one console, full control.
Not tied to any single AI provider. Administrators choose which models power which features — Claude for code, Gemini for docs, swap anytime. A model router handles the abstraction so changing providers is a config change, not a code change.
Built on a Git provider abstraction layer. GitLab today, GitHub or Bitbucket tomorrow. The same interface covers repos, branches, commits, MRs, issues, and labels across any platform.
Workspace administrators manage users, model configuration, integration connections, and cost limits. Full visibility into AI spend, execution history, and team activity from a dedicated admin area.