# Changelog of AI Token Saver — Deterministic Tools for AI Agents (`michiganpublictech/ai-token-saver`) Actor

- **URL**: https://apify.com/michiganpublictech/ai-token-saver/changelog.md
- **Full Actor documentation**: https://apify.com/michiganpublictech/ai-token-saver.md

## Changelog — AI Token Saver — Deterministic Tools for AI Agents

### 0.1.0 — unreleased

Initial release.

- 86 deterministic operations in 9 families: web & HTML, text & tokens, extraction, PII & safety, JSON & data,
  math & units, date & time, validation, encoding & utilities. Full reference in OPERATIONS.md.
- Single, batch (`inputs`) and pipeline (`pipeline`, with `$input` / `$prev` / `$steps[N].path` references) modes.
- Lean dataset items with per-item token-savings estimates (`stats`) and a run-level `SAVINGS` record;
  configurable `llmPricing`.
- Pay per event: `operation` ($0.001) and `page-fetch` ($0.001, successful fetches only). Failed operations,
  `list_operations` and `explain_operation` are free; spending limits are checked before each operation.
- Actor Standby support: HTTP endpoints `GET /`, `GET /operations`, `GET /operations/<name>`, `POST /run`.
- Safety: size caps, ReDoS-guarded user regexes, XML DTD/entity rejection, YAML alias-bomb guard, AST-whitelisted
  expressions, SSRF-guarded fetching, robots.txt (RFC 9309), prompt-injection and PII/secret hints (ATS rules).