01Bottom-up adoption
You told the team to move fast. They did.
The best teams adopted AI tools without waiting for permission, and that was the right call. Now nobody can say what you pay in total, who owns each tool, or what it's producing.
The ops agent for your AI & SaaS tools
Your team's new toolbox, finally under control.
Tally is the ops agent that finds every paid tool, tracks seats and tokens live, and shows you what the spend actually produced.
Built for the tools your team actually uses
Live
- OpenRouter
- Claude Code
- Codex
- Slack
Beta
- BetaChatGPT
- BetaClaude Code
- BetaCodex
- BetaCursor
- Coming nextLovable
- Coming nextVercel
- Coming nextFramer
- Coming nextWebflow
- Coming nextFigma
- Coming nextCanva
- Coming nextMidjourney
- Coming nextAdobe Firefly
- Coming nextNotion
- Coming nextmonday.com
- Coming nextLinear
- Coming nextSlack AI
- Coming nextHubSpot
- Coming nextChatGPT Team
- Coming nextClaude admin data
- Coming nextGemini
- Coming nextSlack
- Coming nextEmail
02The dual economy
You're paying twice, in two different currencies.
Seats are billed every month, used or not. Tokens and credits cost whatever you used, in units every vendor defines differently. Both land on the same card with no link to who used them or what they made.
Seats
Notion, Canva, Figma, ChatGPT Team, HubSpot, monday.com, Linear, Slack AI, Framer, Webflow, and Lovable and Vercel team plans.
The waste: idle seats, forgotten trials, and two or three tools doing the same job across teams.
Tokens & credits
Lovable credits, Vercel usage, Midjourney and Adobe Firefly generations, ChatGPT and Claude API calls, and AI credits inside work tools.
The waste: runaway loops, mid-month credit exhaustion, and personal card top-ups nobody decided on.
- Month-end invoice
- Unvalidated total
- Which team?
- Unknown
- Which artifact was produced?
- Unknown
- Who approved the top-up?
- Unknown
Example
$4,860no owner, no output
- Seats
- $2,940
- Tokens & credits
- $1,920
03How it works
It doesn't wait for you to look. It tells you.
Tally is not another dashboard. It is an ops agent: it checks your tools and bills continuously and comes to you, in Slack or email, with specific findings and a next step.
How the check works
Billed · Invoices
10
Claude Team, 10 seats, $300
Reported · Provider admin APIs
10
seats assigned
Used · The desktop app
6
people active this month
Validated view
4
idle seats to review before renewal
An illustration: invoices are planned, the figures are examples.
- Tally for MacSynced 3 min ago
- Today
- $4.20
- This week
- $18.50
- This month
- $61.03
- Claude Code and Codex tracked
- Tool scan: 7 tools found on this Mac
The desktop app lives in your menu bar. Example figures. 
Sample data: the demo team's budgets this month.
Integrations
Built for the tools your team actually uses.
Live
Model gateway
OpenRouter & model APIsWhat Tally watches
Token spend by team, tool and model; caps on gateway traffic
- Live
Claude Code, Codex, Cursor
Claude Code, Codex, CursorWhat Tally watches
Seats and token usage (Claude Code & Codex Live/Beta; Cursor scan Beta)
- LiveCLI (Claude Code, Codex)
- BetaTally for Mac (Claude Code, Codex), Mac app scan (Cursor)
Where findings go
Slack, emailWhat Tally watches
Budget alerts today; agent findings next
- LiveBudget alerts (Slack, Email)
- Coming nextFindings (Slack, Email)
Beta
Chat & models
ChatGPT, Claude, GeminiWhat Tally watches
Team seats vs. active users; API token spend (desktop apps found by Mac app scan)
- BetaMac app scan (ChatGPT)
- Coming nextChatGPT Team, Claude admin data, OpenAI admin data, Gemini
Coming next
Build & ship
Lovable + Vercel, Framer, WebflowWhat Tally watches
Plans and credits, Vercel usage, sites and apps that went live
- Coming next
Design & creative
Figma, Canva, Midjourney, Adobe FireflyWhat Tally watches
Editor seats vs. active editors, overlapping tools, generations used
- Coming next
Docs & work
Notion, monday.com, Linear, Slack AIWhat Tally watches
Seats, AI add-ons and credits, docs and projects created
- Coming next
Go-to-market
HubSpotWhat Tally watches
Seats by hub, AI credits, campaigns sent
- Coming next
04Artifact attribution
Spend next to what it made.
Tally connects each tool's cost to the artifacts your team produced.
For founders and operators: adopt fast, stay in control.
Founders & COOs
One answer to "what are we paying for, and what did it produce?"
- Hard caps and budget alerts in Slack and email
Next
- Total visibility into the company AI toolbox
- Proactive findings on idle seats and duplicate tools
- Artifact attribution: delivered sites, apps and docs linked to spend
Founders & COOs
One answer to "what are we paying for, and what did it produce?"
- Hard caps and budget alerts in Slack and email
Next
- Total visibility into the company AI toolbox
- Proactive findings on idle seats and duplicate tools
- Artifact attribution: delivered sites, apps and docs linked to spend
Team leads
See your team's tools, seats and what they shipped
- Per-project token and model usage breakdown
- Usage quotas without slowing down creative work
- Near-real-time team activity and session history
Next
- Deliverables and output connected to tool costs
Finance
Every paid tool, explained before renewal
- Monthly budget alerts before overages occur
- Read-only finance role with company-wide visibility
Next
- Fixed seats and consumption tokens side by side in dollars
- Card subscription matching and chargeback exports
The people doing the work
Freedom to use the tools, with nothing to hide
- Background sync with zero prompt or file capture
- Own sessions and project totals
- Personal statusline and menu-bar popover
Next
- Seats and credits next to your own output
05Privacy boundary
We measure tools and output, not people.
Reports are team-level by default. No screenshots, no keystrokes, no rankings.
Stays on the computer
- Prompts and responses
- Tool calls and their output
- File contents and paths
- Source code
Sent to Tally
- Token counts per model
- Timestamps and message count
- Tool and session id
- Project label and branch
The gateway never stores what passes through it.
Stays on the computer
- Prompts and responses
- Tool calls and their output
- File contents and paths
- Source code
Sent to Tally
- Token counts per model
- Timestamps and message count
- Tool and session id
- Project label and branch
The gateway never stores what passes through it.
Every team its own tenant
Teams are isolated from each other. Every query is scoped to the caller's team. The isolation is covered by automated tests.
Secrets encrypted
Provider keys and webhooks are encrypted. Each secret is bound to its own team. Admin keys are read-only.
Adoption, not surveillance
Team-level insights and streaks. Nobody is ranked by tokens or spend. People see their own usage and team totals.
FAQ
Straight answers.
Is this another dashboard?
No. Tally is an ops agent. It checks your tools and bills continuously and comes to you, in Slack or email, with specific findings and a next step. Budget alerts work today; findings on idle seats, duplicate tools and spend without output are coming next.
Does Tally monitor my employees?
No. We measure tools, spend and output, not people. Prompts, responses, code and files stay on the computer. Reports are team-level by default; nobody is ranked.
Which tools does it work with?
Today: Claude Code and Codex (CLI, and the Tally Mac app in beta), OpenRouter and any tool routed through the Tally gateway, with alerts in Slack and email. The Mac app (beta) also finds AI tools installed on each computer. Coming next: Lovable and Vercel, Figma, Canva, Notion, ChatGPT Team, HubSpot, monday.com, Linear, Slack AI, Midjourney, Adobe Firefly, Framer and Webflow, plus Claude and OpenAI admin-panel spend.
Is there a Mac app?
Yes. Tally for Mac is in beta (v0.1.1, Apple Silicon and Intel). It lives in the menu bar, pairs your computer, finds your AI tools and syncs usage in the background. Install it from the download page (access is currently gated to invited teams; requires GitHub sign-in with repo access, or ask your team admin for the installer). Prefer the terminal? Developers can also connect Claude Code and Codex using the Tally CLI (
tallyless).How do I get access?
Tally is in internal beta. If your team has been invited, sign in with your work Google account. If not, request access at hello@ai6labs.com and we'll add you to the beta.
Why are seats and tokens counted separately?
Many tools are both: a plan per person plus a credit pool that runs out.
How does the seat check work?
Billed: AI subscription invoices from the admin's email, or uploaded by hand. Reported: read-only admin keys, OpenRouter today, Claude and OpenAI next. Used: what each person actually uses, measured on their own computer. The three are checked against each other.
What's the difference between equivalent API value and actual spend?
Equivalent API value: what subscription use would cost at list prices. Actual spend: money really paid. The two are never added together.
Where does our data live, and who can see it?
Each team is its own tenant, isolated from every other team, and provider keys are encrypted and bound to their team. The details are in the Privacy Policy.
Tally is in internal beta.
We're running Tally with a small group of teams first. Have an invite? Sign in with your work Google account. Want in? Request access and tell us which tools your team runs on. For founders and operators of teams of 20–150.