TokenFence Now Has a Free Tier — And Here's Why We're Giving Away 95% of the Product

The Developer Tool Pricing Problem

Most developer tools get pricing wrong in one of two ways:

1. Gate everything behind a paywall. Nobody tries it. Nobody recommends it. You get zero organic growth.
2. Give everything away for free. You get 10,000 users and zero revenue. Your investors start asking hard questions.

We just shipped TokenFence v0.3.1 (Python) and v0.2.1 (Node.js) with a third approach: give away the entire core product, charge for the operational layer.

What's Free (Everything That Matters)

TokenFence Community Edition includes:

• Budget enforcement — hard caps, soft limits, per-workflow budgets
• Auto-downgrade — swap to cheaper models when approaching limits
• Kill switch — stop, warn, or raise when budget is exceeded
• AgentGuard Policy engine — full least-privilege enforcement (allow/deny/require_approval)
• Audit trail — every tool call logged with timestamps and context
• OpenAI + Anthropic support — wrap any compatible client
• Async support — full async/await in both Python and Node.js

That's not a demo. That's the full product. No call limits, no feature gates, no trial expiration.

What's Pro ($29/mo)

Pro unlocks the operational layer — the things you need when you're running agents in production and your CFO starts asking questions:

• Real-time cost dashboard — see spend across all agents, all workflows, live
• Slack & webhook budget alerts — get notified before budgets blow, not after
• Multi-agent budget pooling — shared budgets across agent fleets with per-agent sub-limits
• Priority support — direct line to the team when production is on fire

How It Works in Code

Install and use immediately — zero configuration needed:

from tokenfence import guard
import openai

# Works immediately — Community Edition, fully functional
client = guard(openai.OpenAI(), budget="$1.00", fallback="gpt-4o-mini")

# After 50 API calls, you'll see a one-line upgrade nudge in logs
# It never blocks, never degrades, never nags excessively

# When you're ready for Pro features:
from tokenfence import set_api_key
set_api_key("tf_your_key_here")  # Dashboard, alerts, pooling unlocked

Same in Node.js / TypeScript:

import { guard, setApiKey } from "tokenfence";
import OpenAI from "openai";

// Fully functional immediately
const client = guard(new OpenAI(), { budget: "$1.00", fallback: "gpt-4o-mini" });

// Upgrade when ready
setApiKey("tf_your_key_here");

The Soft Nudge — Not a Nag

After 50 tracked API calls, TokenFence logs a single upgrade message. Then it stays quiet for another 100 calls. Then another gentle reminder. That's it.

No pop-ups. No degraded performance. No "trial expired" blocking screens. No passive-aggressive "you've used 90% of your free quota" emails. Just a respectful note in your logs that there's more available if you want it.

Why This Model Works for Developer Tools

The math is simple:

• Free users are your marketing team. Every developer who installs TokenFence via pip install tokenfence is a potential advocate. They'll mention it in code reviews, recommend it in Slack, star it on GitHub.
• Production users have real budgets. When you're running 10,000 agent calls/day and your AI bill is $15K/month, $29/mo for a dashboard that shows where every dollar goes is a no-brainer.
• The conversion trigger is natural. You don't need to convince anyone to pay. When they need visibility into production costs, they'll find the Pro features themselves.

Get Started

# Python
pip install tokenfence

# Node.js
npm install tokenfence

Full docs at tokenfence.dev/docs. Questions? Open an issue on GitHub.

TokenFence — budget caps, kill switches, and least-privilege policies for AI agents. Free forever. Pro when you're ready.