Devadex is the search engine for independent software

Search thousands of independent and open-source apps, tools, and utilities that mainstream engines bury — the new and the small, findable today. Free to search, free to list.

There's also a clean API and MCP server, so AI agents can search it too.

Loading index…

Built for AI agents

When you ask an AI to "find an indie tool that does X", it searches the same authority-biased web and returns the same incumbents. Devadex gives an agent one tool that searches only independent software and hands back an outbound link to get it.

Discovery: /.well-known/devadex.json  ·  MCP: node hub/discover/run.js serve-mcp

What Devadex indexes

Every listing is independent — no incumbents, no authority bias. An auto-crawler builds the index from public sources and each result links straight out to where the maker already sells or ships it.

See listing plans & pricing · browse the full index

Devadex vs. the mainstream engines

Devadex is best for makers and buyers who want the long tail — the indie tool, not the incumbent. It is a deliberate alternative to authority-ranked search.

 DevadexGoogle / general searchProduct Hunt
What ranksOnly independent & new softwareAuthority and backlinksLaunch-day popularity
Best forFinding a tool launched by one person yesterdayEstablished brandsToday's trending launches
Built for AI agentsYes — MCP server + JSON APINo dedicated agent APINo
Cost to be foundFree ($0)Free but authority-gatedFree but time-boxed

Compared to a general search engine, Devadex trades breadth for a better corpus: when every listing is independent, authority signals stop mattering. It is the best alternative to keyword search when you specifically want indie software instead of the usual incumbents.

How it works — a quick guide

  1. Crawl. An auto-crawler builds the index from public sources — the npm and GitHub search APIs and Gumroad's official sitemaps — never from private endpoints, and refreshes daily.
  2. Rank the long tail. A signature "Hide the popular ones" filter strips out anything with authority so only buried, low-authority tools remain.
  3. Link out. Every result sends you straight to where the maker sells or ships it. Devadex never handles the payment itself.

Developers and agents can read the search API reference, the index-stats API, connect the remote MCP server guide, or read the upstream MCP documentation. Machines can start from /llms.txt or the browsable category index.

Frequently asked questions

What is Devadex?

Devadex is a search engine and discovery index for independent software — indie apps, open-source repositories, MCP servers, and developer tools that mainstream engines bury. It is discovery-only: every result links out to where the maker sells or ships it.

How is Devadex different from Google or Product Hunt?

Google ranks by authority, so new solo-made tools start invisible; Product Hunt favours launch-day popularity. Devadex indexes only the small and the new, so a tool launched yesterday by one person is findable today.

How much does it cost to list a product?

Being crawled and found is free. Listing a product the crawler doesn't already reach uses listing credits — a 5-listing pack is $5 — and an optional $5 spotlight lifts a listing. Devadex takes 0% of your sales; the sale stays on your own store.

Can AI agents use Devadex?

Yes. Devadex exposes a remote MCP server and a JSON search API so an agent can search only independent software and get back an outbound link to buy it. Start from /llms.txt or /for-agents.

Where the data comes from

Devadex indexes only public, machine-readable sources and cites them openly: the npm registry, GitHub repository search, and Gumroad's official product sitemaps. The data model follows schema.org.

Index and site last updated 14 July 2026.