LedgerLens — a personal ledger for AI subscriptions and quotas

One page for all your AI coding subscriptions and usage, so you can see what's wasted

Problem

The signals are dense: one user agonizes over switching from Claude to ChatGPT for coding, another burns through the company Copilot quota and is left idle, a third gets their Claude org disabled and has to fall back to Codex, and someone else watches a third-party relay site return 502 and vanish with their money. Heavy users hold multiple AI subscriptions plus metered APIs and have no tool telling them what the month actually cost, which quotas went unused, or which service quietly raised prices.

Solution

A local-first AI spend ledger. It ingests API usage exports and CSVs, tracks fixed subscriptions manually or via import, and normalizes everything into cost-per-task by editor, model, and agent. It flags unused quota, idle subscriptions, overlapping tools, and recent price hikes, then sends a weekly actionable note like "merge these two low-use subscriptions" or "move this task class to a cheaper model."

Users

Indie devs and 2–10 person teams paying for Claude, ChatGPT, Cursor, Copilot, and relays; engineering managers who reimburse or budget for the team.

MVP

Manual entry and CSV import for subscriptions; usage export ingestion for two major APIs; monthly cost and cost-per-task charts by tool and model; unused-quota and overlapping-subscription flags; exportable monthly report for reimbursement.

Revenue

Free tier capped at 2 subscriptions; $5/month or one-time unlock for unlimited tracking and reports; per-seat add-on for small teams.

Why now

AI subscriptions and APIs are both ballooning, each vendor bills in its own way, relays can vanish overnight, and individuals need their own books — exactly the small painful problem incumbents ignore.