Ai Spend Agent

aibill — financial accountability for AI agents

[](https://github.com/futurastudio/ai-spend-agent/actions/workflows/ci.yml) [](https://www.npmjs.com/package/ai-spend-agent) [](LICENSE) [](package.json)

**Know what cost evidence exists for your AI agents, what drove it, and what to do next—with the source and limits attached.**

npx aibill         # short form — same CLI as `npx ai-spend-agent`

[Documentation](https://asktilden.com/docs) · [Supported sources](https://asktilden.com/docs/sources) · [MCP setup](https://asktilden.com/docs/mcp) · [Roadmap](https://asktilden.com/docs/roadmap)

Building a Node integration? The [supported library preview](docs/LIBRARY.md) documents the narrow, packed-test contract for usage-record validation, analysis, Receipt v0, pinned standards projections, and terminal rendering.

**Version boundary:** the npm `latest` release (live version in the badge above) reads supported Claude Code, Codex, and experimental Gemini CLI financial evidence. Gemini remains `fixture_verified`, financial-only, and excluded from statusline, Glance, Context Health, plan/runway, invocation evidence, recommendations, and Apply.

**The guided token-reduction and local accountability loop is published.** Run `npx aibill improve` from the project you want to improve — or practice safely first with `npx aibill improve --sample`, a labeled demo that writes nothing. The guided flow gives one personalized test, tracks matched before/after completed-session snapshots, and prints a percentage only when the user's declared quality held. An MCP client can draft the plan conversationally and hand you one `npx aibill improve --draft …` command; APPROVE is st