ThinkWatch Lite

Overview

ThinkWatch Lite is a local gateway for AI coding clients such as Claude Code and Codex. It runs on macOS, Windows, and Linux, letting developers switch upstream providers without reconfiguring clients, keep API keys away from relays, block malicious tool calls, and trace every request. Written in TypeScript, it is a community-maintained observability tool.

Best for

Best for
Developers using Claude Code or Codex who want local control over upstreams and request visibility.

Use cases

• Switch between upstream AI providers without editing client configs
• Protect API keys from third-party relay services
• Block dangerous tool calls and trace request activity

Notes

ThinkWatch Lite is a local gateway for AI coding clients such as Claude Code and Codex. It runs on macOS, Windows, and Linux, letting developers switch upstream providers without reconfiguring clients, keep API keys away from relays, block malicious tool calls, and trace every request. Written in TypeScript, it is a community-maintained observability tool.

1,054 stars on GitHub. Last updated 2026-10-05. Licensed MIT.

Use cases

• Switch between upstream AI providers without editing client configs
• Protect API keys from third-party relay services
• Block dangerous tool calls and trace request activity

Pros

• Cross-platform local gateway
• Keeps API keys local and out of relays
• Provides request tracing and tool-call blocking

Cons

• Requires running and maintaining a local service
• Community project with limited support
• Adds a configuration layer for AI clients

Indexed from awesome-llmops and enriched against its public facts.