VibeCheck vs. the provider consoles
Provider consoles report what you spent with that provider. VibeCheck, Maestro AI's coding agent visibility product, shows what the spend bought: every Claude Code, Codex and Cursor session with its engineer, repo, kind of work and any pull request it fed. Most teams keep the console for billing and use VibeCheck to explain the bill.
Start free for up to 5 engineers. Team views come with Pro.
What each one answers
| Question | Provider consoles | VibeCheck |
|---|---|---|
| How much did we spend? | Yes. The invoice is the source of truth. | An estimate, priced from each session's own tokens at published reference prices |
| Who spent it? | Varies. Anthropic's Claude Code dashboard lists spend per user; Cursor's lists usage per member. | By person and team, across tools |
| On what kind of work? | Not described in their docs | Feature, bugfix, debugging, refactor, infra, review and explore |
| What did a pull request cost? | Anthropic's Team and Enterprise dashboard counts PRs written with Claude Code. None describes cost per PR. | Cost per authored PR, with the GitHub App connected |
| What happened in an expensive session? | Activity counts, not session detail | Who ran it, what it was for, and whether it planned, tested, reviewed or needed rework |
| All coding agents in one place? | One provider each | Claude Code, Codex and Cursor (desktop agent) together |
| Does it match the invoice? | Yes | No. Labeled as an estimate; discounts and subscription plans aren't applied |
Console column from each provider’s public docs, November 2026: Claude Code analytics · Cursor team analytics · Codex governance
Keep the console for billing
The console is the source of truth for what you owe. VibeCheck doesn't replace it and won't match it to the cent. It answers the next question finance asks: what was this for?
How VibeCheck gets the session view
One install command per laptop, or a rollout through your MDM. A stop hook in Claude Code, Codex or Cursor nudges a small background service and returns at once; it never waits on the network. Because capture happens on the machine, it works whether Claude Code runs through Anthropic directly or Amazon Bedrock.
Questions
- Does VibeCheck replace the Anthropic or OpenAI console?
- No. Keep the console for billing. VibeCheck breaks the spend down by person, team, kind of work and pull request.
- Why don't the numbers match our invoice?
- Each session is priced from its own tokens, cache included, at current published reference prices. Negotiated discounts and subscription plans aren't applied, and every figure is labeled as an estimate.
- Which tools does it cover?
- Claude Code, Codex and the Cursor desktop agent, through Anthropic directly or Amazon Bedrock.
- Does it slow our engineers down?
- No. The hook returns immediately; uploads happen in the background and retry if a laptop is offline.
See what your agent spend bought.
Start free for up to 5 engineers. Team views come with Pro.