Unlimited seats.
Pay for the loops you run.
Credits pay for the loops that run on your code — not for the people on your team. Business runs on frontier models from Anthropic and OpenAI. AI Sovereignty keeps every AI request in Europe. Every plan includes unlimited seats, workspaces, and repos.
Trial
3 loop runs on 1 repository
See what the loops find. Card required; cancel before day 3 and pay nothing.
- ✓On Business or AI Sovereignty, your pick
- ✓Converts to that plan when it ends
- ✓Pay now any time to lift the limits
- ✓Unlimited seats
- ✓All sixteen loops, switched on per workspace
- ✓Proposals you greenlight or decline
- ✓MCP server for your coding agent
Business
€468 billed annually
≈ 50 loop runs / month
Frontier models from Anthropic and OpenAI.
AI Sovereignty
€564 billed annually
≈ 50 loop runs / month
European AI. Nothing goes to a US AI provider.
- ✓Loops and chat on GLM 5.3, an open-weight model
- ✓Served in the EU by Melious and Mistral
- ✓No AI request ever goes to Anthropic, OpenAI or any US provider
- ✓Unlimited seats, workspaces and repos
- ✓All sixteen loops, on every repo you connect
- ✓Credit usage breakdown and on-demand top-ups
- ✓Priority queue
- ✓Email support
Unlimited seats, workspaces, and repos. Plan credits refresh monthly; top-up credits carry over. Loop runs are an estimate for one repository: a run that finds nothing costs less, a deep pass more. Annual billing saves 20%. Prices exclude VAT, added at checkout.
Credits
Pay for what runs, not who's on the team.
Credits measure the AI compute a loop run or a request actually uses — not the number of people in your workspace.
~1 credit
A quick answer about your code
~2–3 credits
A spec you write with Tekk by hand
~1–3 credits
One loop run — a run that finds nothing costs less than one that drafts a proposal
Run out mid-sprint?
Chat keeps working and bills €0.60 per extra credit. Loops pause until your credits refresh or you top up, so nothing runs up a bill unattended. You get warned at 80%.
FAQ
Common questions.
What's the difference between Business and AI Sovereignty?
Which AI does the work. Business runs your loops and chat on frontier models from Anthropic and OpenAI. AI Sovereignty runs everything on GLM 5.3, an open-weight model, served in the EU by Melious and Mistral: no AI request ever goes to a US provider, and the plan locks that in for the whole workspace. Both sell the same credits and the same sixteen loops. AI Sovereignty is our standard plan: we are a European company, and we think your code should stay on European AI.
Why does AI Sovereignty cost more?
Running open-weight models in Europe has not come out cheaper per loop run than the US frontier models. You get the same credits on both plans; AI Sovereignty is priced about 20% higher.
What's a credit?
Credits measure the real AI compute each piece of work uses. A loop run — a loop waking, reading what changed and bringing you a proposal or an honest "nothing found" — is charged for the work it actually did, so a run on a small change costs less than a deep pass over a whole area. A loop run averages about 2 credits; a quick answer is about 1; a spec you write with Tekk by hand 2–3. The amount scales with the real work, not by user or session.
What does a loop cost when it finds nothing?
Less than one that finds something. A run that reads the change, opens its one skill and finds nothing worth raising stops there and reports an honest negative — it doesn't draft a proposal, so it uses fewer credits. You see every run, empty or not, in your credit usage breakdown.
How do I control what the loops cost?
Loops are opt-in per workspace: switch on the ones you care about and leave the rest off. A loop wakes when your code changes, when a connected tool like Sentry raises a signal, or when part of its area has gone unexamined for a while — so a quiet repository runs fewer loops. Every run shows up in your credit usage breakdown.
What happens when I hit my monthly limit?
Chat keeps working: extra credits bill at €0.60 each on your next invoice. Loops and other background work pause at zero and pick up again when your credits refresh or you buy a top-up — nothing unattended adds to your bill. You'll get an in-app warning at 80% usage.
Can I add credits without upgrading my plan?
Yes. Subscribers can buy top-up packs in-app anytime: 50 credits for €29, 150 for €79, or 400 for €209. They're billed once and added immediately to your balance.
Do unused credits roll over?
Plan credits don't — your plan's balance refreshes each billing cycle. Top-up credits never expire: they sit alongside your plan credits, are used after them, and carry over until spent.
How many seats are included?
Unlimited, on every plan. Add your whole team, contractors, or stakeholders — there's no per-seat charge. Credits are shared across the workspace.
Can I change how many credits I get, or switch plans?
Any time. Both plans let you pick your monthly credit volume — from 100 up to 1600 — and each step up doubles your credits. You can also move between Business and AI Sovereignty and keep the same credits. Change it whenever your workload does; the new amount applies from your next billing cycle.
Is there an annual discount?
Yes — 20% off when you pay annually, at any credit volume. Toggle to annual billing above to see the discounted rate. You're billed once for the year.
Can I cancel anytime?
Yes. No contracts, no cancellation fees. Cancel from your billing settings and your plan stays active until the end of the current billing period.
See what the loops find in your code.
Three days, three loop runs on your repository. Card required, cancel any time. Connect your repo and switch the loops on.
npx @tekk_coach/cli@latest initForClaude Code, Codex or Cursor
Set up Tekk in this repository: run npx @tekk_coach/cli@latest init in the repo root and follow its prompts. When it finishes, run /ideate-loops to pick the loops for this project.Run it in your repo. It signs you in through the browser, connects the repository, and wires Tekk into Claude Code, Codex and Cursor. It shows every file it will change and asks once. No AI runs during setup, so it costs no credits. Agent setup docs →