Navigation
Articles
- Agent-Friendly Web
- Depth-First DevRel
- Internetless Software Development
- Moloch / Complexity as Strategy
- Outcomputed
- Reciprocal Networks of One-Person Companies
- Scalable Knowledge Work
- Spec-First Development
- The World Is Information
- Biscuit
- Blixt
- Claude Code Routines
- Cloudflare
- Cloudflare Artifacts
- Code From Anywhere
- Foreside
- GitHub Agentic Workflows
- Iterate
- Minion.ai
- Moyai
- New Nexus
- Parallel AI
- Product Hunt
- ProfitPath
- Slimmer AI
- Sprent
- Topliner
- Artem Murzin
- Ben Allfree
- Ben Tossell
- Dmitry Alexenko
- EastlondonDev
- Enes Akar
- Guillaume Lebedel
- Ives van Hoorne
- Jan Wilmake
- Jonas Templestein
- Kristian Freeman
- Matt Carey
- Maurice Kleine
- Paraga Agarwal
- Ricky Robinett
- Ryan Hoover
- Sam Goodwin
- Soham Ganatra
- Sunil Pai
- Thomas Slabbers
- Travis Fischer
openapisearch.com
openapisearch.com
openapisearch.com is an API discovery and search tool developed by Jan Wilmake under Code From Anywhere, derived from ActionSchema's OpenAPI tooling. It allows developers and AI agents to search for APIs and retrieve their OpenAPI specifications. The tool went viral in March 2025.[1]
Origin
openapisearch.com grew out of Wilmake's work on ActionSchema's OpenAPI proxy and search infrastructure during 2024. ActionSchema needed a way to discover and query external APIs from within data pipelines; the underlying search component was extracted and made into a standalone product.[1]
A key product principle: the search results deliberately do not expose full information about the sorting algorithm or authentication details, protecting Wilmake's data advantage while still providing useful output to end users.[2]
Viral Traction (March 2025)
In March 2025, openapisearch.com experienced a significant viral event alongside related tools in the forgithub ecosystem, driving substantial traffic and developer attention. The launch was coordinated with other tools: zipobject (repository packaging), sponsorflare (billing), and various forgithub utilities.[1]
Relationship to Other Projects
openapisearch.com sits within a broader tooling ecosystem:[1]
- ActionSchema — parent project and conceptual origin
- uithub — sister tool for repository context (code); openapisearch handles the API layer
- forgithub family — tools for interacting with GitHub data programmatically
- Parallel AI — Wilmake later extended API search and MCP work at Parallel AI
Technical Notes
- Deployed on Cloudflare Workers (edge compute)
- Results served from a proprietary providers dataset (
providers.json) kept behind a private key - Search outputs are rate-limited and deliberately incomplete to protect the moat[2]
January 2026: SLOP and agentskills.io
On January 28, 2026, Jan shipped updates to openapisearch introducing two new features:
- OpenAPI → agentskills.io: A new
/skills/{hostname}endpoint that generates a skill from any OpenAPI spec[3] - OpenAPI → llms.txt (SLOP): Converting any OpenAPI spec into a Simple Language Open Protocol (SLOP) document that is llms.txt compatible — reducing GitHub's 2M-token OpenAPI spec to ~30k tokens[3]
Jan coined the term SLOP (Simple Language Open Protocol) for the human-readable, llms.txt-compatible format distilled from OpenAPI specs. Every SLOP can be used via the LLMTEXT MCP.
February 2026: Cloudflare Code Mode MCP
In February 2026, Cloudflare launched a "Code Mode MCP" which converts OpenAPI specs into code-mode style tool calls without a traditional MCP server layer. Jan's response: "For big APIs like this, this beats CLIs! MCP is back. Inspires me to do MORE with my openapi-mcp-server."[4] The Cloudflare approach validated Jan's earlier OpenAPI-to-MCP direction while showing a convergent path in the ecosystem.