Back to directory
Docs & Writingteamlatten/@latten/cost
@latten/cost
Cut LLM costs in your AI app and coding agents. On-device detection of wasted tokens (redundant calls, stuck loops, oversized + uncached context) with copy-paste fixes. Local-first, values-free.
Suggested install command
npm view @latten/costAlways inspect the linked repository and skill instructions before running commands. Skills are instructions; permissions and execution still matter.
Compatibility
Agent support matrix
3 supported
| Agent | Status |
|---|---|
| Claude Code | Supported |
| OpenCode | Not listed |
| Cursor | Supported |
| MCP | Not listed |
| GitHub Copilot | Not listed |
| Windsurf |