scruple.io

What It Is

Scruple is an AI spend monitoring tool built for engineering organizations, positioned as an “attribution ledger for AI usage.” It does not provide large language model capabilities itself. Instead, engineers launch AI coding tools such as Claude Code, Cursor, and Codex through the Scruple CLI, which records each session’s token spend, model mix, duration, repository, branch, ticket, engineer, and git activity. The data is then aggregated by project, PR, ticket, feature, or team.

Core Capabilities

Its key value is tying costs directly to work. Compared with the built-in dashboards from Anthropic, OpenAI, or Cursor, which typically show spending by API key, seat, or overall usage, Scruple aims to answer more granular questions: how much AI budget a specific feature, PR, or ticket consumed, and which team, repository, model, or tag is causing budget leakage. The site also mentions People view, Tags view, and Analytics view, which can be used to see engineer-level spending rankings, team/project rollups, and trend-based forecasts for next month’s bill.

Integrations, Pricing, and Privacy

Scruple is tied into the engineering workflow through its CLI and explicitly supports Linear and Jira integrations, allowing AI sessions to be linked to tickets. At present, the official website only offers a waitlist and does not disclose pricing, free quotas, deployment options, API availability, SSO, access controls, or enterprise compliance information. Because the tool handles sensitive engineering metadata such as repositories, branches, tickets, engineer identities, and git activity, companies should carefully verify the scope of data collection, whether code content is uploaded, data retention policies, permission isolation, and security certifications before procurement.

Pros, Cons, and Best-Fit Users

Its main strength is its very focused positioning. It is well suited to engineering teams that are already using AI coding tools at scale but lack cost attribution capabilities. For CTOs, engineering productivity teams, and FinOps teams, it can help with budget governance, cost forecasting, and identifying high-value usage patterns. The limitations are also clear: the product still appears to be in a waitlist stage, so its maturity is unknown; it depends on engineers using tools through the CLI, which requires workflow adoption; and there is currently no information about Chinese-language support, payment methods, or access from mainland China.

Access from China and Alternatives

Network accessibility from mainland China is unknown, and payment options have not been specified. If Scruple is not usable, teams can first consider the built-in admin dashboards from OpenAI, Anthropic, or Cursor, or build cost collection and attribution through an internal AI gateway, logging system, Langfuse, Helicone, Portkey, and similar tools. However, these alternatives usually require more in-house integration work and may not directly provide the PR-, ticket-, and feature-level cost ledger that Scruple emphasizes.