Cut your AI bill.
Keep your prompts.
PromptLite is a tiny overlay that lives above every AI composer your team uses. It shows token cost in real time, then optimises long prompts in one tap — without changing meaning.
142tokensWorks everywhere your team prompts
From bloated prompt to shipped reply in seconds.
Install in 2 minutes
Browser extension, VS Code plugin, or `npm i -g promptlite`. No account required for the free tier.
Type anywhere
PromptLite watches your composer on Claude, ChatGPT, Gemini, Lovable, and your editor — and shows token cost in real time.
Accept the optimised prompt
Cross your budget? PromptLite suggests a tighter version, shows a diff, and replaces it in one tap. Meaning preserved.
Six surfaces. One overlay.
Each card opens a working PromptLite simulation on that platform — no install needed.
47% fewer tokens on long research threads.
Compress sprawling context blocks before they hit Anthropic's pricing.
Pin the extension — PromptLite mounts inside claude.ai automatically.
Live budget badge inside the composer.
Watch cost tick up as you type. One-tap rewrite before you ever press send.
Works on chatgpt.com and the desktop app webview.
Diff preview without leaving Google's UI.
Side-by-side before/after with annotated edits and undo.
Also available on Firefox and Edge.
Stop blowing your Agent-mode budget.
Critical warnings before a run, attachment-aware optimisation built in.
Detects Agent-mode runs and warns before you burn credits.
Inline cost in Copilot & Claude panels.
Token-aware ghost text right where you ship code.
Lights up in every chat panel: Copilot, Claude Code, Cursor.
`promptlite "…"` — scriptable, CI-friendly.
Pipe prompts through stdin, get optimised output and stats back.
Drop-in for any terminal LLM CLI. Use as a pre-commit hook to fail builds on CRITICAL prompt issues.
Everything you need to control AI spend.
Real-time token budgets
Per-user, per-team, per-project. Soft & hard caps with Slack alerts. CFOs finally get a number.
Diff preview
See exactly what changed before you accept. Tag-annotated edits.
Budget alerts
Warn at 80%. Block at 100%.
Attachments aware
Handles files.
Multi-surface
Same overlay on Claude, ChatGPT, Gemini, Lovable, VS Code and CLI.
Privacy-first
On-device tokenisation. Prompts never leave your browser unless you opt in.
A CFO-grade dashboard for your AI stack.
Roll up token usage by team, project, model and surface. Set budgets, see who's burning, export to finance. Built for orgs serious about AI ROI.
Open the dashboardToken usage · This month
$12,847 savedPay for the seats — not the tokens.
Free
For solo devs trying it out.
- Browser extension
- CLI tool
- Token badge on all surfaces
- Local optimisation
Pro
For power users & freelancers.
- Everything in Free
- Cloud-backed history
- Diff annotations
- Priority models
- Usage analytics
Team
For teams that ship with AI.
- Everything in Pro
- Shared budgets
- Slack alerts
- Admin dashboard
- SSO
Enterprise
Procurement-friendly.
- Everything in Team
- On-prem option
- Custom models
- SOC 2 / DPA
- 24/7 support
Real teams. Real savings.
"PromptLite cut our Claude bill by 38% in the first month. Nobody changed how they work — the badge just nudges you."
"Finally a tool a CFO can read. Per-team budgets, hard caps, and a dashboard that doesn't lie."
"I run it in CI through the npm package. Long prompts get rewritten before they ever ship to prod."
Questions, answered.
Stop overpaying for tokens.
Install PromptLite in two minutes. Free forever for solo devs.