For everyone who works with AI$50 per person

Make AI-powered work easier to plan, manage and account for.

Works besideClaude CodeCursorCodexGemini CLI
Checkout project · AI spendExample

With VerGen, every task is estimated before the work and accounted for after.

One square is $0.25 of AI spend

Works with the tools already in use

Whichever AI tool does the work:Claude CodeIn the terminalCursorIn the editorCodexIn the terminalGemini CLIIn the terminal
  • TrackersJiraLinearAsanaClickUpTrellomonday.comAzure DevOps
  • CodeGitHubGitLabBitbucket
  • ChatSlackMicrosoft Teams
  • HelpdesksZendeskIntercomHubSpot

Before the work

Plan.

Every task gets a cost estimate before it starts.

Large tasks are broken into steps, each with only the context it needs. The estimate is posted as one note in the ticket, and kept up to date.

VerGen Agent · on your machineExample
  1. Read PAY-142 from your tracker

    Add SSO to the admin · Checkout

  2. Sized only the files the ticket changes

    No model called · every file it counted is listed

  3. Estimated $3.40

    High confidence
  4. Planned 4 steps, adding up to $3.40

    1. auth/sso.ts$0.90
    2. admin/login.tsx$0.60
    3. tests$0.40
    4. iteration$1.50

    Each step carries only the context it needs

  5. Note posted to PAY-142 in your tracker

PAY-142 · in your trackerTo do

Add SSO to the admin

Checkout

VerGen · one note, kept up to date

  1. EstimatedHigh confidence$3.40
  2. Planned4 steps$3.40
  3. SpentAfter the work—
Fig. 01 — The VerGen Agent estimates and plans PAY-142, then posts one note to the ticket

In the console

Manage.

One place to see what AI work cost.

Spend against the estimate, task by task. What each AI tool cost, at your own rates. And a word where the team already talks when a week leaves its usual band.

VerGen · CheckoutExample

Projects

  • Checkout
  • Mobile app
  • Support replies
  • Data pipeline
128 of 141
tasks estimated before the work
112
within their estimate
13
started with no estimate

Spend against the estimate, task by task

EstimateSpentOver

What each AI tool cost

Claude Code$1,840
Cursor$960
Codex$410

At your own rates

VerGenAPPSlack · Teams

Checkout's week left its usual band: review took 2.1 days, against a usual 0.8.

PAY-131Over estimate

Move invoices to the new tax table

Estimate $6.20Spent $7.90

Recorded, not stopped. A person decides what happens next.

Fig. 02 — The Checkout project in the console, with the notes it sends

After the work

Account for.

What it cost and what was saved.

For the managers who answer for the spend: each project's retrospective, and what to change next.

Beside your tools, never in their way.

How it sits with your work

i.

Runs no coding agent

Your people keep their own AI tools, and VerGen holds none of their credentials.

ii.

Reads only what it sizes

Every estimate lists the files it counted.

iii.

Records, never blocks

Work that runs past its estimate is recorded, not stopped. A person decides what next.

iv.

Never bills by token

One price per person. No invoice carries a token figure.

Questions

What people ask first.

Who is VerGen for?

Everyone who works with AI, not only engineers: they get a cost estimate and a plan for each task before the work starts. And the managers who answer for the spend, who read each project's retrospective: what the work cost, what was saved and what to change next.

How is a saving measured?

Against the same project's tasks of the same size that ran with no estimate: saved is what estimated tasks cost below them. Lost is what un-estimated tasks cost beyond what VerGen estimates for their size. Both are priced at the vendor's list price from recorded sessions, and too few tasks to compare shows as not comparable, never a guess.

Do I have to use your coding agent?

There is none. VerGen runs no coding agent of its own: your people keep Claude Code, Cursor, Codex or Gemini on their own machines, and the VerGen Agent beside them reports each session. VerGen estimates the ticket before the work and records what the VerGen Agent pushes.

How is it priced?

One plan, $50 per person a month, with the VerGen Agent included for everyone. There is no base fee and no limit on people, teams or projects, and invoices never carry a token figure.

Which tools does it work with?

Jira, Linear, Asana, ClickUp, Trello, GitHub Issues, Azure DevOps and monday.com for tickets; Zendesk, Intercom, Freshdesk and HubSpot for helpdesks; GitHub, GitLab, Bitbucket and Azure Repos for code; Slack and Teams for notifications.

Does it see what people do on their own machines?

Only on machines enrolled in the workbench. Each session's turns, tokens and setup are recorded and scored against your team's briefs; the text is kept for 90 days, the scores for good. Local work is never billed.

5 more questions
How is this different from an agent orchestration platform?

Orchestration platforms optimise how much your agents get through. VerGen estimates each ticket before the work starts and records what it cost. You can run both.

What happens if the work needs more than its estimate?

It is recorded, not stopped. When the change merges, the estimate is set beside what every session on the ticket sent, and the ticket shows whether it landed within it. Every band names where it was made: in VerGen or on the machine.

Who can estimate a ticket?

Anyone in the project allowed to request an estimate, while the ticket is still a draft. Once a session or a pull request names the ticket, its estimate is fixed; estimating again before then supersedes the last one.

What code does VerGen read?

To estimate a ticket it reads only the files it selects from the repositories the ticket changes, to size them; no model is called. The estimate lists exactly which files it counted.

Does VerGen stop or approve an agent's work?

No. VerGen runs no coding agent and holds none of their credentials. It estimates before the work and records what was done.

Know what AI work will cost, and what it saved.

$50 per person a monthVerGen Agent included for everyoneNo base fee