SettleGrid Help

AI-powered support

How can I help?

Ask me anything about SettleGrid

Glossary

Key terms and definitions for AI service settlement, metering, billing, and the agent payment ecosystem.

AI Service Settlement

The process of tracking, billing, and distributing payment for individual AI service calls across providers, consumers, and intermediaries. Settlement transforms every API call, tool invocation, or agent action into a monetizable event with real-time metering and automatic payouts.

Metering

Real-time measurement of AI service usage at the individual call level. SettleGrid meters every invocation in under 50 milliseconds using atomic Redis operations, tracking call count, token usage, byte transfer, or execution time depending on the pricing model.

Per-Call Billing

A pricing model where consumers are charged for each individual service invocation rather than through subscriptions or prepaid credits. Per-call billing aligns costs with actual usage and is the most common model for AI tools and APIs.

Progressive Take Rate

A revenue-sharing model where the platform fee increases gradually with revenue brackets rather than applying a flat percentage. SettleGrid charges 0% on the first $1,000/month, 2% on $1K-$10K, 2.5% on $10K-$50K, and 5% above $50K. This means developers always keep more than they would on a flat-rate platform.

Smart Proxy

A reverse proxy that adds billing, metering, and access control to any existing API endpoint without code changes. Developers point their API URL at the Smart Proxy, configure pricing, and SettleGrid handles authentication, balance checks, and usage tracking transparently.

Multi-Hop Settlement

Automatic payment distribution across a chain of AI agents or services where Agent A calls Agent B which calls Agent C. SettleGrid tracks the full call chain and distributes revenue to each participant based on their contribution, supporting up to 10 hops per session.

Agent Identity (KYA)

Know Your Agent (KYA) is a verification framework for AI agents that establishes identity, capabilities, and trust level before allowing financial transactions. KYA enables per-agent budgets, spending controls, and audit trails for agent-to-agent commerce.

Cost-Based Routing

An optimization strategy that automatically selects the cheapest AI service provider meeting quality thresholds for a given request. Cost-based routing compares price, latency, and reliability across registered providers and routes the call to the optimal endpoint.

Fallback Chain

An ordered list of alternative AI service providers that are tried in sequence when the primary provider is unavailable or fails. SettleGrid automatically manages failover, retries, and billing across the fallback chain without consumer-side code changes.

Settlement Pulse

A real-time heartbeat of settlement activity across the SettleGrid network. The settlement pulse tracks transaction volume, revenue flow, active tools, and network health, providing both operators and developers with live observability into the AI economy.

Tool Discovery

The process by which AI agents and consumers find, evaluate, and connect to available AI services. SettleGrid provides a Discovery API, a public showcase, and machine-readable service cards that enable programmatic discovery across categories, pricing models, and protocols.

MCP (Model Context Protocol)

An open protocol originally created by Anthropic that standardizes how AI models interact with external tools and data sources. MCP defines a structured interface for tool discovery, invocation, and response handling. SettleGrid adds a billing and settlement layer on top of MCP.

Protocol-Agnostic

An architecture that works across multiple payment and communication protocols without requiring protocol-specific code. SettleGrid is protocol-agnostic, supporting 15 protocols (MCP, x402, AP2, MPP, Visa TAP, UCP, ACP, Mastercard Agent Pay, Circle, REST) through a single unified SDK.

Take Rate

The percentage of revenue that a platform retains from transactions processed through it. Traditional marketplaces charge 10-30% flat. SettleGrid uses a progressive take rate starting at 0% on the first $1,000/month of developer revenue.

Outcome-Based Billing

A pricing model where the consumer pays based on the result or outcome of the AI service rather than the computation performed. For example, paying only when an AI successfully extracts the requested data or completes a task. SettleGrid supports outcome-based billing as one of six pricing models.

Budget Controller

A real-time spending limit system that prevents AI agents or consumers from exceeding predetermined budgets. SettleGrid enforces budgets at the per-call level with atomic balance checks, ensuring agents cannot overspend even under concurrent request load.

Agent Firewall

A security layer that validates, rate-limits, and filters AI agent requests before they reach the underlying service. The agent firewall combines KYA identity verification, budget enforcement, rate limiting, and anomaly detection to protect both providers and consumers.

Ready to build with SettleGrid?

Start with the free tier. 50,000 operations per month, progressive take rate starting at 0%.

Get Started Free Read the Docs