Skip to content
2026-06-23

500+ tools and counting

Correction (2026-10-02): this post called every tool deterministic. The utility tools are; the catalog also holds model-backed tools (the /v1 model gateway, report products, media and answer tools), each marked modelBacked on /api/pricing.

The Agent402 catalog passed the 500-tool mark. Each tool is priced per call and callable with a single HTTP request, and every tool CI can run without a third-party key is tested against its own example on every CI run. Here's a look at what's in the box.

Note (updated 2026-08-18): this post originally counted every catalog entry, including hundreds of near-duplicate pairwise converters that were later collapsed into a handful of parameterized tools. The catalog is quoted as an evergreen "500+ tools" everywhere now; the exact live number is always at /api/pricing and /health.

What categories exist

Updated 2026-10-11: the list below is rewritten to name only tools in the catalog today; several tools this post first listed have since been retired.

The catalog spans 20+ categories, grouped into "kits" - each kit is a focused collection of related tools:

  • Data processing: JSON (format, validate, query, diff, merge, flatten), CSV, XML and YAML conversion and validation
  • Web tools: rendering, extraction, link listing, sitemap parsing, screenshots
  • Search: web search, news, images, suggestions, and cited answers (the answer tool is model-backed)
  • Finance: stock quotes and history, SEC filings, XBRL company facts, insider trades
  • Crypto & DeFi: token prices, TVL, wallet balances, ENS resolution, gas prices
  • Government data: FRED economic series, Treasury rates and the yield curve, CPI and unemployment
  • PDF processing: PDF to markdown, text extraction, page extraction, metadata
  • Media: image conversion and resizing, OCR, audio transcription (model-backed)
  • Barcode & QR: QR generation, barcode decoding and product lookup
  • Security: DNS lookup, TLS certificate info, WHOIS, HTTP headers, SPF checks
  • Encoding: base64, hex, URL encoding, number base conversion
  • Math & stats: statistical summaries, correlation, linear regression, GCD/LCM
  • Hashing: SHA-256, SHA-512, MD5, HMAC, checksums
  • Text & string: diff, Levenshtein distance, keyword extraction, case conversion, text stats
  • Agent memory: wallet-keyed persistent storage (write, read, recall, forget, log, grants)

How we got here

We started with 50 tools in the first week. The approach was simple: pick a category, build 5-10 tools that cover the common tasks, write the CI test for each, and ship. Each kit follows the same pattern - a single file that exports an array of tool definitions with handlers.

The constraint that kept quality high: every tool CI can run without spending upstream must answer its own example input correctly in CI. If a tool can't pass that bar, it doesn't ship. (The metered tools, which need third-party keys, are covered by input probes instead.)

What's coming next

We're continuing to expand the catalog based on what agents actually request (tracked via the /api/find endpoint). The most-searched-for capabilities that don't yet have tools get built first. Current priorities include more data transformation tools, additional financial data sources, and deeper government data coverage.

The catalog is also listed in the x402 Bazaar discovery listing, where any x402-compatible agent can find its tools.

Back to blog