Get Started

Vektra Lite

The open-source version of Vektra - self-hosted, free, MIT licensed

Vektra Lite is the open-source version of Vektra — the same demo-intake and support-ticket workflow, self-hosted by you. One Python process, one Postgres database, no subscriptions, no dashboards to babysit. MIT licensed — take it, run it, fork it.

The source lives on GitLab: gitlab.com/dapmedia/Vektra-Lite

What's included

🎵 Demo intake — artists run /submit, fill a small form, and the track lands as a review card in your private staff channel with its own discussion thread.
⚖️ Team decisions — Approve, Reject (with a reason), or DM from the card. The artist is notified by DM instantly.
🎟️ Support tickets — members open a ticket from a button; staff get a card and thread, and members can reply by simply DMing the bot.
🧑‍🎤 Artist self-service — /my_submissions and /my_stats so artists can check where they stand.
🧹 Staff tooling — /queue, /recent, and /submission for a clean overview, all ephemeral.
🤖 Built-in AI assistant (MCP) — set one password and connect Claude Desktop, claude.ai, Cursor, or any MCP client. The AI can review demos, manage tickets, and create custom slash commands that go live instantly.
🧠 Optional AI spam screening — flag suspicious submissions with one AI key (Groq, OpenAI, or SambaNova).
🎶 Optional music — /play through your own Lavalink node, behind a master switch.

Get the sourceClone the repository and read the full setup guide on GitLab.Compare with hosted VektraSee what the paid tiers add on top of the open-source core.

Requirements

Need	Details
Python	3.10 or newer
Database	Any Postgres — Neon, Supabase, or your own
Hosting	Runs on a ~150 MB free VPS
Discord bot token	From the Discord Developer Portal

Setup in four steps

Create the Discord bot

Make an application in the Discord Developer Portal, reset the bot token, and enable the Message Content Intent. Invite the bot with bot + applications.commands scopes. This server becomes the bot's permanent home — Vektra Lite is single-server by design.

Grab a free Postgres URL

Neon, Supabase, and self-hosted Postgres all work. Use sslmode=require with any managed provider.

Configure and install

Copy .env.example to .env, fill in your bot token, DATABASE_URL, and (optionally) MCP_PASSWORD for the AI assistant. Then install the dependencies from requirements.txt.

Run the bot

Start bot.py. It creates its own tables on first boot — no manual migration needed — and the staff setup commands guide you through the rest from inside Discord.

Lite vs hosted Vektra

	Vektra Lite	Vektra (hosted)
Price	Free, forever (MIT)	Free tier + paid plans
Hosting	Your VPS	Managed by us
Servers per bot	1 (single-server lock)	Unlimited
Web dashboard	❌	✅ (Pro)
White-label bot identity	❌	✅ (Pro+)
Managed storage	Your database	Included
Updates	Pull from GitLab	Automatic

FAQ

Is it really free?

Yes. Vektra Lite is MIT licensed with no paid tier, no seat limits, and no usage caps. You host it, you own it.

Can multiple servers use one Vektra Lite bot?

No — the first server to invite the bot claims it forever, and any other server is left automatically. Set GUILD_ID to pin a specific server instead.

Does the AI assistant cost extra?

The MCP server is built in, but it needs an AI client of your own (Claude, Cursor, etc.) and the bot's optional features use whichever AI key you provide. The bot itself adds no markup.

How is this different from Vektra's free plan?

Vektra's free plan is hosted — we run the infrastructure and you get updates instantly. Vektra Lite is the self-hosted core: you run everything, and in exchange you give up the dashboard, multi-server support, and managed storage.

Where do I report bugs or ask for help?

Open an issue on the GitLab repository. Contributions are welcome.