01OBSERVED ACTIVITY
Know where the tokens go.
Filter reported input, output, and cache usage by repository, project, agent, model, member, and period. Missing counters stay unknown.
A CLEARER VIEW OF AGENT WORK
Every session
leaves a footprint.
See the work behind your coding agents. Follow tokens, sessions, and prompts from local logs to a shared repository view.
A free service by BornWest. Your prompt content is a choice.
LOCAL SIGNAL → SHARED CONTEXT
CodexSession records
Claude CodeLocal activity
More agentsSupported sources
Your collectorPreview · scope · share
Repository workspace
repository / project
01 Sessions & user turns
02 Reported token usage
03 Permitted prompt history
GitHub access · explicit privacy
01 / THE WORKSPACEUnderstand the activity.
Understand the activity.
Improve the practice.
Start with what the sources report. Use those records to make the next conversation, review, or repository workflow more useful.
02SOURCE RECORDS
Follow a session.
Move from a repository to its sessions and eligible user prompts. Keep agent, model, timestamps, and source provenance in view.
03REVIEWABLE SUGGESTIONS
Give the next prompt more shape.
Review suggestions derived from stored activity. Keep scope, expected results, and verification clear, with links back to the evidence.
04DRAFT SKILLS
Make a useful pattern reusable.
Prepare repository skill drafts with source sessions and evaluation steps. Review and copy the Markdown when it is ready to use.
05TEAM PLAYBOOKS
Turn lessons into team practice.
Build scoped playbook drafts and onboarding exercises from observed work. A person reviews examples before sharing or adoption.
06PLANNING EVIDENCE
Plan with a record to point to.
Use evidence summaries to frame next steps. Separate reported activity from assumptions about effort, outcomes, or future demand.
Workflow drafts use deterministic templates over permitted stored telemetry. They do not inspect repository code, call a model, or install themselves.
02 / HOW COLLECTION WORKSOn your machine.
On your machine.
In your control.
The collector runs beside the agents you use. The workspace helps you inspect what those supported sources actually recorded.
- 01 / CONNECT
Sign in with GitHub.
Choose your personal or organization workspace. Repository visibility follows verified GitHub access; admins manage their workspace.
- 02 / PREVIEW
Choose what leaves your Mac.
Enroll the local collector, choose a prompt mode, and inspect the preview. Collection starts from enrollment; history needs a separate backfill.
- 03 / UNDERSTAND
Explore the work together.
Open the private dashboard for usage, session details, and drafts. Local queues retain events while the collector is offline.
Full prompts require your opt-in.
Counts-only mode sends no prompt bodies. Redacted mode uses best-effort removal. Full mode shares eligible prompt text with permitted workspace viewers. Prompts can contain private information; preview before starting. The collector does not read source-code files or record your screen.
03 / KNOW THE BOUNDARIESA useful view starts
A useful view starts
with honest coverage.
Installed does not mean captured. Heartbeat, first activity, source support, and measured usage are different signals.
Codex · Claude CodeSupported local session records
OpenCode · Pi · GrokAdapter support; source format matters
AmpExplicit stream capture required
Remote developmentCollector on that host, or supported export
Other applicationsNo universal prompt interception
04 / GET STARTEDYour next session.
Your next session.
A clearer record.
Install on each Mac where your agents run. Then sign in, choose your workspace, preview, and start the collector.
macOS · Python 3.9+ · Git
Node.js 18+ for the npx setup below.
Open workspace setup Node.js 18+ for the npx setup below.
Loading this server’s installation command…
Installation is configured by this FootPrint server.
NEXT
Open workspace setup to generate commands for your account, repository scope, and prompt choice.
footprint doctorcheck source coveragefootprint previewinspect before sharingfootprint statusinspect collector healthInstall without Node.js
Loading…
Downloads a checksum-verified CLI into your user account. No sudo. Inspect the installer before running it.
The record is a start.
The learning is the point.
Cross-agent counters are useful. The opportunity is to connect trustworthy activity to repository context, reviewed recommendations, and reusable team practices.
That loop is a product hypothesis. FootPrint is an early, free service; better outcomes need evidence from the teams who use it.
Built by BornWest06 / OPEN PRODUCT THINKINGExplore the business.
Explore the business.
Keep assumptions visible.
FootPrint is free to use. These public experiments explore a hypothetical business model, independently of your workspace data.
A
Revenue arithmetic
Assumptions, not a forecast.
Hypothetical MRR$19,900
Annualized, unchanged$238,800
Organizations × monthly USD = MRR. Annualized = MRR × 12, not recognized annual revenue. Excludes churn, acquisition, discounts, costs, and taxes.
B
12-month scenario forecast
An editable model, not a demand prediction.
Month 12 MRR—
12-month revenue—
Revenue less fixed costs—
Inspect monthly calculations
| Month | Organizations | Revenue |
|---|
Each month: prior organizations × (1 − churn) + new organizations. Revenue uses that end-of-month count, including fractional expected organizations. No cohort timing, variable costs, acquisition, staffing, discounts, or taxes. Independent of calculator A.
Adapted from the FootPrint market review, October 17, 2026. Defaults are editable hypotheses. No customer, revenue, or market-size claim is implied.