Know what your AI actually gets you.
See what Claude Code, Cursor and Copilot actually produced: the features, fixes and PRs behind every session. Then learn what your best people do differently, so the whole team can do it too.
Free for one person · Set up in a minute · Prompts stay on your laptop
Works with the tools you already use
- Claude Code
- Cursor
- GitHub Copilot
- ChatGPT Enterprise
- Claude.ai
- Codex CLI
- Gemini CLI
- AWS Bedrock
- Google Vertex
- Azure OpenAI
- Windsurf
- Devin
Running in a minute. Nothing to change about how you work.
- 1
Install it
One command on your laptop. No admin rights, no proxy, nothing in front of your AI tools.
curl -fsSL https://metermark.ai/install | sh - 2
Keep working
It reads the history your AI tools already keep, labels each session by the kind of work, and syncs every 15 minutes.
- 3
Get answers
Open your dashboard, or just ask Claude: Metermark comes with an MCP server and a read-only API.
Start with yourself. Roll it out when it’s useful.
The same install works for one developer or ten thousand. Pick where you are.
Know what your AI habit is really costing you.
Claude Code, Cursor, Copilot: five tools, five bills, no idea which sessions were worth it. Metermark puts them on one page, split by what the work actually was.
- One total across every AI tool you use
- Spend split by kind of work: features, bug fixes, research
- Commits and lines changed next to what they cost
Find out who’s great at this, and make everyone that good.
Some people get far more out of the same tools. Metermark shows you who, what they do differently, and writes it into the CLAUDE.md and AGENTS.md files your team’s tools already read.
- Compare teams and projects side by side
- Your best people’s habits, turned into rules for everyone
- Mark a project a win and see exactly how it happened
Every AI tool, every team, one answer for the board.
Push it to every laptop through your MDM in an afternoon. Then answer “what are we getting for this?” with real numbers instead of seat counts.
- Signed .pkg and .msi for Jamf, Kandji, Mosyle and Intune
- SSO, SCIM, an audit log and a signed DPA
- A read-only API and MCP server for your BI tools and board deck
We never see the conversation.
One small row of labels and counts leaves your laptop per session. There’s nothing for us to leak, and nothing for anyone to read over your shoulder.
What you type
Labelled on your machine, then thrown away. The label can’t quote you.
What the AI writes back
Never opened. We note the model and the cost, nothing else.
What your code does
Commits and lines changed are counted locally. Only the count leaves.
Team numbers by default. Naming a person takes the right role and leaves an audit trail. Read the privacy policy
Simple pricing.
Free for one person. $20 a seat per month for a team. That’s the whole price list.
Just me
Free
$0 one person, your own data
- Your AI spend, split by kind of work
- 50 labelled sessions a month
- Coverage map and privacy receipt
My team
Team
$20 per seat / month, billed monthly
- Every session labelled, no limit
- One seat per person, nothing else to buy
- Projects, team comparisons, custom connections
My company
Enterprise
Custom one annual contract for the whole organization
- SSO, SCIM, audit log, API export
- Connections to your internal tools
- Signed DPA and a person to call
Frequently asked questions
What is Metermark?
Metermark shows engineering teams what their AI coding spend is worth: cost and output per tool, who uses it best, and which projects drove the value.
Which AI tools does Metermark cover?
Claude Code, Cursor, GitHub Copilot, ChatGPT Enterprise, Claude.ai, Codex CLI, Gemini CLI, AWS Bedrock, Google Vertex, Azure OpenAI, Windsurf and Devin, each labelled by how much the vendor exposes. It measures across the tools your company pays for, not one vendor dashboard.
Does Metermark store conversation content?
No. Not a word of the conversation is stored: not the prompt, not the reply, not what a tool returned. The prompt is labelled on your machine and thrown away; one row of labels and counts per session leaves it.
Who is Metermark for?
Engineering leaders (VP Engineering, Platform, DevEx) who have to show what AI coding tools are worth, and the finance or procurement people who own the AI tool budget. One developer can use it on their own data.
Who makes Metermark?
Hassion Studio, a product studio. The maker page is at hassion.studio/makes/metermark.
How much does Metermark cost?
Free for one person with their own data, 50 labelled sessions a month. Team is $20 per seat per month, billed monthly. Enterprise is one annual contract. Start free on metermark.ai.
See your own numbers today.
Make an account, run one command, and your dashboard fills in as you work.
curl -fsSL https://metermark.ai/install | sh