Pricing
One price per seat. You pay your providers directly.
CodeVector is a flat per-developer-seat license on an annual contract. No percentage on top of your tokens, no markup on usage. Your AI bill stays yours.
How pricing works
Flat per-seat license
One annual price per developer. Your quote depends on one number: how many seats you need.
0 percent take-rate
We take nothing from your token spend. Provider rates pass through untouched.
Built for 10 to 500 developers
The best fit is 25 to 150 seats. Smaller team? Ask us and we will tell you honestly whether it is worth it yet.
What every seat includes
Cost control
- Per-seat budget caps, daily, weekly, or monthly
- Per-IDE and per-repo cost attribution, built in
- Rate limits on requests, tokens, and concurrency
Access and security
- Access grants per user and per group
- Secret-pattern blocking before a prompt leaves your network
- Synchronous audit log on every admin change
Routing and tooling
- Multi-provider routing, including open-weight models
- Setup recipes for Claude Code, Cursor, Continue, Zed, aider, and opencode
- One-command developer onboarding with the codevector CLI
Support
- A support contract that covers provider drift
Every item is documented: see budget caps, cost attribution, secret blocking, and the CLI.
Three ways to pay for AI coding, side by side
Uber spent its 2026 AI budget in about four months, then capped every engineer at $1,500 a month per tool. A blunt cap is one answer. Here is how the options compare.
| Personal seats | Blunt monthly cap | CodeVector | |
|---|---|---|---|
| Cost visibility | None across the team | A ceiling, no breakdown | Per developer and per repo |
| Effect on heavy users | Unbounded spend | Hard stop mid-task | Budget plus routing to cheaper models |
| Shadow usage | Unmeasured, on expense reports | Still invisible | One endpoint, all of it logged |
| What finance gets | A surprise | A number, no story | Chargeback by team and repo |
Try it on a hosted pilot
Tell us about your team on the contact form and we will reach out within one business day to set you up with a hosted CodeVector instance. Nothing to install: point Claude Code or Cursor at it and watch the real per-developer and per-repo spend appear. Pilots usually run one to two weeks, longer if your evaluation needs it. When you are ready for production, you self-host on your own infrastructure.
Questions about pricing
- How much does CodeVector cost?
- CodeVector is a flat per-developer-seat license on an annual contract. The quote depends on one number: how many seats you need. There is no surcharge on your token spend; you pay your AI providers directly at their published rates. Contact us for a quote for your team size.
- Is there a per-token surcharge?
- No. CodeVector takes 0 percent of your token spend. You pay the provider, and a flat per-seat license to CodeVector. Your bill stays predictable as usage grows.
- How does the pilot work?
- Submit the contact form with a few details about your team. We reach out within one business day and set you up with a hosted CodeVector instance, nothing to install. Pilots usually run one to two weeks, longer if your evaluation needs it.
- Do we self-host in production?
- Yes. In production CodeVector runs on your own infrastructure, on any cloud or on-prem, in a container. The hosted pilot is there so you can try it quickly before you deploy your own.
- What team size is a good fit?
- Teams of 10 to 500 developers, with the best fit between 25 and 150. If you are below 25, contact us and we will tell you honestly whether it is worth it yet.