价迹·AI工具定价数据集

# PriceTrail / 价迹 — AI & SaaS Pricing Dataset (derived layer) **EN.** A continuously-updated open dataset of **AI / SaaS pricing snapshots and pricing-change timelines**. Each row in `data/price_cards.jsonl` is a structured pricing observation (tool, vendor, tier, price, billing cycle, source, capture time); each row in `data/price_changes.jsonl` is a detected pricing change event between two snapshots (price up/down, tier added/removed, free-tier cut, billing cycle change, ...). The dataset is regenerated from live pricing pages on a fixed cadence, so it doubles as a **pricing time-series** for competitive intelligence, procurement, VC research, and model training. **中文.** 一份持续更新的开源「AI / SaaS 定价」数据集，包含定价快照卡与定价变更时间线。 `data/price_cards.jsonl` 每行是一条结构化定价观测（工具、厂商、档位、价格、计费周期、 来源、抓取时间）；`data/price_changes.jsonl` 每行是两次快照之间检测到的定价变更事件 （涨价 / 降价 / 新增档位 / 下架档位 / 取消免费档 / 计费周期变更 …）。数据集按固定节奏从 真实定价页重新生成，天然是一份「定价时间序列」，可用于竞品情报、采购、VC 研究与模型训练。 > **Freshness / 新鲜度:** regenerated on every run (default cadence: every 12h); > the dataset always reflects the most recent observed pricing. ## License / 许可 - The **derived pricing observations** (structured price points, change events, provenance) are released under **CC-BY 4.0** — see `LICENSE`. - Source pricing pages retain their own terms; attribution to each source is preserved via `source_url`. ## Non-affiliation / 非附属声明 This dataset is **independent and not affiliated with, endorsed by, or sponsored by** any vendor, product, or service named within it. All product names, logos, and trademarks are the **property of their respective owners** and are used here only to identify the pricing subject of each observation. Prices are observed facts redistributed under CC-BY 4.0; this dataset makes no claim of partnership with any company. ## Coverage honesty / 覆盖诚实声明 This release covers the **publicly-parseable pricing pages** that the PriceTrail pipeline can fetch and parse (AI / SaaS / DevOps / collaboration tools, etc.). A number of vendors gate their pricing pages behind bots / login walls (e.g. some LLM providers and design suites); those are **not yet** in this dataset and are explicitly excluded — this dataset does **not** claim full vendor coverage. See the per-run source list for the exact coverage of this build. ## Fields / 字段 **price_card** (13 fields): id, tool, vendor, tier, price, currency, billing_cycle, effective_date, source_url, snapshot_ref, captured_at, lang, category. See `schema.json` → `definitions.PriceCard`. An optional `category` groups tools by type (e.g. `agent` for AI-agent apps such as Cursor / WorkBuddy / Coze); rows without a category are omitted (legacy sources). **price_change** (8 fields): id, tool, change_type, before, after, detected_at, evidence_snapshot_before, evidence_snapshot_after. See `schema.json` → `definitions.PriceChange`. **change_type** vocabulary: - `new` - `price_up` - `price_down` - `free_tier_cut` - `feature_cut` - `tier_added` - `tier_removed` - `cycle_change` ## Load / 加载示例 ```python from datasets import load_dataset cards = load_dataset("Myco-Net/pricetra-pricing", data_files="data/price_cards.jsonl")["train"] changes = load_dataset("Myco-Net/pricetra-pricing", data_files="data/price_changes.jsonl")["train"] ``` ## Stats / 本版统计 - price_card rows: 13107 - price_change rows: 13