Products

nasikolabs 推出 Open Runtime 平台,为 AI Agent 提供逐 token 成本追踪

Why we picked it

nasikolabs 出了个 Open Runtime,装一条命令就能给现有 Agent 做 token 计量,不用改代码。能看到每个 MCP 工具、每次运行烧了多少钱,到预算上限还会自动把 Agent 暂停。跑多供应商 Agent 团队的工程负责人应该看看。

nasikolabs 发布了一款 Open Runtime 平台,用于解决 Agent 花费分散在各个供应商账单上、没有按 Agent 细分的问题。平台以一条命令安装,位于团队现有 Agent 之下运行,不需要在代码里引入任何 SDK。

Body

安装后,它可以把 token 消耗按团队、Agent、模型和会话维度拆分,覆盖所有供应商,粒度可下探到单个 MCP 工具和单次运行。预算控制方面,消耗达到预算 80% 时触发告警,到达上限时 Agent 自动暂停。

对同时使用多家模型供应商的团队来说,这补上了 Agent 场景下成本可见性缺失的一环,也把硬性预算上限从人工盯账单变成运行时自动执行。

Original · rohanpaul_ai

Most teams run the token race blind, since agent spend lands on separate provider invoices with no per-agent breakdown.

@nasikolabs launched an Open Runtime platform that sits beneath existing AI agents and prices every token they burn.

installs with one command and sits under the agents a team already runs, so no SDK goes into anyone’s code.

From there it splits spend by team, agent, model and session across every vendor, down to individual MCP tools and runs.

At 80% of budget, an alert fires. At the limit, the agent pauses.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Products" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include nasikolabs, MCP and tools, agents, token cost.

Source attribution: curated from rohanpaul_ai (@rohanpaul_ai).