Vibestrate costs $0. It is free and open source under the Apache 2.0 licence, with no paid tier, no seat pricing, and no usage fees. Your only cost is whatever the AI model vendors charge you directly, which using Vibestrate does not change.
What you get
| What | Vibestrate |
|---|---|
| Price | $0 |
| Licence | Apache 2.0, including commercial use and a patent grant |
| Included | Everything: CLI, terminal UI, Mission Control dashboard, orchestrator, flows, all eleven providers |
| Limits | None. No seat cap, no run cap, no feature gate |
| Account required | None. There is no sign-up and no billing relationship |
| Runs on | macOS, Linux, Windows (WSL) |
Why there is nothing to charge for
Vibestrate is never in the middle of a model call. It spawns the provider CLIs you already installed and already authenticated, and reads their output. Prompts and responses travel straight between that CLI and its vendor. There is no proxy to meter, no tokens to resell, and no hosted service to fund - so there is no margin to take and no reason to gate a feature behind a plan.
That is also why your API keys never reach Vibestrate. They stay in the vendor CLIs that already had them.
What you pay the model vendors
Whatever they already charge you. Vibestrate adds no markup and takes no cut, so your Anthropic, OpenAI, or Google bill is exactly what it would be if you ran those CLIs by hand. What changes is that you can see the cost: every run records an estimated token and dollar figure locally, computed on your machine from cached public pricing tables.
That ledger is an estimate for budgeting, not billing data. The vendor's own dashboard remains the authority.
A crew made entirely of local models through Ollama, LM Studio, or llama.cpp costs nothing at all. Mixing is normal: an expensive model on the plan, a cheap one on the build, a different frontier model on the review.
For AI agents
A machine-readable version of this page lives at /pricing.md, in the same format as llms.txt. Both say what this page says.
