Letta · FinOps Profile

Letta Ai Finops

FinOps view of Letta spend. Letta Cloud (Developer API plan) bills a monthly base subscription plus metered usage - dollars per active agent per month and dollars per second of tool execution - on top of pay-as-you-go LLM and embedding usage that is passed through from whichever model provider (OpenAI, Anthropic, Google, etc.) an agent is configured to use. The Personal Free and Pro tiers are flat-rate or free, bounded by agent count rather than metered usage. The core Letta server is Apache-2.0 licensed and free to self-host, in which case the dominant costs are your own compute plus whatever model provider API keys you bring.

Letta Ai Finops is the FinOps profile for Letta on the APIs.io network, aligned with the FinOps Foundation Framework.

It defines 4 billable meters, billed in USD, on a monthly cycle, and pricing category usage-based.

The profile maps 8 FOCUS columns for cost-allocation reporting.

Tagged areas include AI, Agents, LLM, Memory, and FinOps.

Category: AI and Machine Learning Pricing: Usage-Based Billing: Monthly FOCUS v1.3
AIAgentsLLMMemoryFinOpsCost ManagementFOCUS

Framework Alignment

Framework
FinOps Foundation Framework
Data Spec
FOCUS v1.3

Charge Categories

UsagePurchaseAdjustment

FOCUS Columns

BillingCurrency
USD
ChargeCategory
Usage
InvoiceIssuerName
Letta
PricingCategory
Usage-Based
ProviderName
Letta
PublisherName
Letta
ServiceCategory
AI and Machine Learning
ServiceName
Letta Cloud API

Meters

active_agents
Unit: agents
Active stateful agents, billed per agent per month on the Developer API plan.
tool_execution_seconds
Unit: seconds
Wall-clock seconds of tool execution across all agents, billed per second.
model_provider_usage
Unit: tokens
Pass-through LLM and embedding token usage billed by the configured model provider, not marked up by Letta.
self_hosted_compute
Unit: hours
Infrastructure cost of running a self-hosted Letta server (no Letta invoice).

Sources