For Individual Developers

Your AI usage, finally visible

You use eight different AI tools a week. You have no idea what any of them cost, which models you're burning money on, or how much time you lose context-switching between them. FORG fixes that — one lightweight agent, complete clarity.

Download FORG See pricing
8
AI tools per week
Average developer juggles 8 different AI coding tools weekly — each with its own cost, model, and session history.
34%
of AI time is overkill
One-third of AI usage goes to expensive frontier models for tasks a cheaper model handles just as well.
15 min
lost per day to context switching
Average daily overhead from jumping between AI tools — that's over 60 hours a year of fragmented flow.
The problem

Flying blind on your own AI bill

As a developer you make dozens of AI-assisted decisions every day. But at the end of the month, you get a single line-item: $140. You have no idea where it went, whether it was worth it, or what you could change.

  1. 01No per-session breakdown — you can't tell if the cost was a 200-line refactor or an autocomplete suggestion.
  2. 02No model visibility — Cursor, Claude Code, and Copilot all call different models. You have no unified view.
  3. 03No efficiency signal — you don't know if switching tasks to GPT-4o-mini would save 60% with identical quality.
  4. 04No history — if your employer audits your AI usage, you have nothing to show except a credit card statement.
forg sessions list --user me --last 7d
$ forg sessions list --user me --last 7d
ADAPTERMODELTOKENSCOSTDURATION
AI coding workspaceAdvanced Balanced142,800$0.4348m
IDE workspaceAdvanced Balanced89,200$0.2731m
AI coding workspaceStandard Fast312,000$0.0822m
IDE workspaceSmall Fast54,100$0.0117m
IDE workspaceFree Hosted28,400$0.009m
5 sessions · 7-day total$0.792h 07m
Run forg sessions export --format csv to download full history

What changes when you can see everything

Without FORG
  • One opaque monthly bill with no breakdown by session, adapter, or task type.
  • No way to know if you used GPT-4o for a task that claude-haiku-3 could have done at 5% of the cost.
  • Context-switching between 4+ tools with no unified history — you repeat yourself constantly.
  • Zero audit trail if your company asks what AI tools you used or how much you spent.
With FORG
  • Every session logged: adapter, model, token count, cost, and duration — searchable in seconds.
  • Efficiency score surfaces which tasks you over-spent on and suggests cheaper model alternatives.
  • One dashboard unifies Claude Code, Cursor, VS Code Copilot, and any other adapter you run.
  • Full exportable history as CSV or JSON — you own the data, forever, regardless of plan.

Everything in one place

FORG runs as a lightweight local agent. No proxy, no added latency — just instrumentation that finally makes your AI habits legible.

Personal Usage Dashboard

Live overview of your token spend, cost, session count, and top models — updated as you work, not at end of month.

Per-session Cost Breakdown

Every session gets its own cost card: adapter, model, input/output token split, and duration. No black boxes.

Model Efficiency Score

FORG computes a task-weighted efficiency score: are you paying for a sports car when you only need a bicycle?

Adapter Comparison

Compare your Claude Code, Cursor, and Copilot sessions side-by-side: cost per task type, model mix, time invested.

Weekly Digest

Every Monday, a personal email: your top 3 insights, biggest spend, efficiency trend, and suggested model swap.

Privacy-first by design

Your individual session data is never visible to your employer. FORG is personal telemetry — it belongs to you.

Solo plan · $19/mo

Know exactly what your AI usage costs

Install the FORG agent in under two minutes. It works alongside every adapter you already use — zero changes to your workflow.

Download FORG