Best Alternatives to shim-cli in 2025

shim-cli is an open-source local hook and proxy that inspects what coding agents like Claude Code, Codex, Copilot CLI, and VS Code send out, flags secrets and personal data, masks them where possible, and reports per-session usage and cost without accounts or telemetry. If you want similar visibility, redaction, or cost tracking with a different approach, these alternatives are worth evaluating.

LiteLLM Proxy

An open-source LLM gateway that sits between your tools and model providers, giving you request logging, spend tracking, and guardrails. It covers cost visibility and centralized control well, though it is heavier to run than a local hook and its secret masking is less focused on coding-agent payloads.

Presidio (Microsoft)

An open-source PII detection and anonymization toolkit that can be wired into your own pipeline to find and mask personal data before it leaves your machine. It is strong on detection breadth but is a library, not a ready-made agent proxy, so you build the integration yourself.

gitleaks

A fast open-source secret scanner for repos and pre-commit hooks. It is excellent at catching credentials in code, but it scans files rather than live agent traffic, so it complements rather than replaces outbound request inspection.

Helicone

A hosted and self-hostable observability layer for LLM apps with request logging, cost dashboards, and caching. It offers richer analytics than shim-cli but typically requires routing through its proxy and an account, which conflicts with shim-cli's no-account, local-only stance.

Langfuse

Langfuse

In Directory

Observability and analytics for LLM apps

An open-source LLM observability platform for tracing prompts, token usage, and costs. It is well suited to teams wanting shared dashboards and evaluation tooling, but it is a larger self-hosted stack and does not specialize in masking secrets from coding-agent sessions.

OpenLLMetry

An open-source observability framework built on OpenTelemetry that instruments LLM calls for traces, token counts, and cost. It integrates with existing observability backends and is a good fit if you already run tracing infrastructure, though it observes rather than redacts.

Claude Code native hooks and settings

Claude Code itself supports hooks and permission settings that can log activity or block certain actions. This is the lowest-friction option for Claude Code users, but it is vendor-specific and lacks cross-agent support, secret detection, and masking that shim-cli provides.

shim-cli's niche is local, account-free inspection and masking of coding-agent traffic with per-session cost reporting. LiteLLM Proxy and Helicone are the closest matches for gateway-level logging and spend tracking, while Presidio and gitleaks cover PII and secret detection in different layers. Langfuse and OpenLLMetry suit teams wanting full observability stacks, and native Claude Code hooks are the simplest option if you only use one agent. Choose based on whether you prioritize local privacy, cross-agent coverage, or shared team dashboards.