Invoices show what AI costs.
TideFig shows what it returns.
TideFig measures the value and return on investment of your AI spend: how much you pay per person, and what you get back from the people using it.
Free forever for individuals · macOS · no credit card
TideFigTokens
Sessions
Requests
Cost at list price
Anthropic
OpenAI
The gap
You can already see what AI costs. You can't see what that spend returns.
Invoices and token dashboards answer "how much?" but never "was it worth it?". Which people and teams turn AI spend into real output? Which burn money without it? Without that link, every budget decision is a guess.
What you have today
Provider invoices, token counters, per-seat license totals.
What TideFig adds
Early access, separate from the desktop app: spend tied to delivered output and ROI per team.
Same company, two lenses
One flat number. No signal.
Spend per person against yield. High bars = high return; short bars = burn.
The product
One product, two altitudes
The free desktop app is the entry point: your own AI usage in plain numbers. The platform turns that visibility into company-wide AI ROI.
Your personal AI week, in plain numbers
A weekly overview of your own AI usage: instant visibility, free forever.
- This calendar week: tokens, sessions, and requests
- Token share by vendor, then model
- Cost at list price. The tooltip says Estimate, unless the saved mode is API key or company cloud. Subscription usage shows as its own figure, labeled "List price, not charged".
- Claude Code and Codex on this Mac
- Plan limits: 5-hour and weekly percent used, with time to reset, for Claude and ChatGPT / Codex. Claude limits show once you turn them on.
Company AI ROI (early access, not part of the desktop app)
Connect AI spend to value over any period: last 30 days vs the prior one.
AI spend
Cost / output
Team ROI
People · ROI score
Evidence Confidence Rating
A · AuditedHypothetical time savings are kept separate from audited financial returns, so every ROI figure survives finance review.
- KPIs with period comparison: spend, cost per output, tokens, team ROI
- Spend concentration: who drives spend and at what yield
- People table: one row per person, click for the full profile
- Spot your own waste: premium models on trivial tasks, runaway agents
Your data stays on this Mac.
TideFig reads the usage files already on this Mac. Your tokens, costs, and paths stay here. The only things that leave this Mac are an anonymous activity ping, update checks, and a price-table refresh.
Early access: ROI by role
Built for office work, not just developers
In the early-access company product, separate from the desktop app, each team gets outcome metrics that match how their work is judged.
Cost & time per completed reconciliation or close
Close the books faster, with the audit trail intact
Cost per reconciliation
Hours per monthly close
AI spend per close
- Accuracy rate measured on every AI-assisted entry
- Audit approval required before figures count as returned value
Top waste signals
The costliest low-value patterns, ranked by what they burn
Early access, not in the desktop app: the company product ranks where money goes in and nothing comes out, with each signal's monthly cost, a suggested fix and the estimated saving.
$4,440
example monthly waste
Premium model used for trivial tasks
Top-tier models answering one-line prompts. Set a cheaper default and keep premium for complex work.
$2,340
per month
6 people
caused by
Agent runs that never produce a result
Autonomous runs that loop, stall or get cancelled: full token burn, zero delivered output.
$1,120
per month
3 people
caused by
Spend on output that gets rejected
Work thrown away: low accept rates mean the tokens were spent and the output wasn't used.
$980
per month
4 people
caused by
Estimated monthly saving: $3,100 with routing defaults
Protect top performers' budgets instead of capping everyone
Spot skills that correlate with higher yield and make them default
Every insight: finding, evidence, action, estimated saving
Who it's for
Built for the people paying for AI, and the people using it
Individual AI users
"What am I actually using, and what does it cost?"
- See your own usage free every week, with no dashboards to configure
- Know this week's tokens, vendor and model share, and cost at list price, with subscription usage shown separately as "List price, not charged".
- Reads the usage files already on your Mac: no account, no credit card
Company leaders (early access, not the desktop app)
Eng, Product and Finance: "justify the spend, know where it pays off"
- Tie every dollar of AI spend to delivered output, per team and per person
- ROI scores and one-line reasons: Good / Watch / Bad, with evidence
- Waste signals ranked by monthly cost, with suggested actions
- One-line recommendations per person: coach, reroute, or protect budget

Start with your own week. Scale to your whole company.
The free desktop app reads the usage files already on your Mac. Company AI ROI is rolling out to early-access teams now.