Skip to content

Refy vs XERJ

Refy is an open-source MCP server for developer tools. XERJ is an open-source MCP server for knowledge and memory.

  • Refy is best for pick the right data table component.
  • XERJ is best for search local code and docs from an agent.
What it is
Refy is an open-source MCP server for developer tools.
XERJ is an open-source MCP server for knowledge and memory.
Best for
Pick the right data table component
Search local code and docs from an agent
Kind
MCP server
MCP server
Pricing
Open source
Open source
Platforms
API, CLI
CLI, Linux, macOS, self-hosted, API
Runs
Remotely
Locally
GitHub stars
19.2K
Last commit
Nov 2026
License
Apache-2.0
Open source
Good for
Pick the right data table component; Verify generated screen code uses real props; Derive a theme from a brand site
Search local code and docs from an agent; Give an agent persistent memory notes; Look up real API definitions

What Refy does

  • Search by need

    Finds a component family from a written need in three languages.

  • API lookup

    Reads a component's documented API.

  • Code validation

    Checks screens against the real API.

  • Theme extraction

    Builds light and dark AA-contrast themes from a site.

  • Plugin shortcuts

    Claude and Codex plugins add screen and theme skills.

What XERJ does

  • Single binary

    Search, vectors, facts and memory in one Rust program.

  • Evidence-backed answers

    Answers come with the program and evidence used.

  • Runs locally

    Works offline on a CPU, on a laptop or server.

  • Elasticsearch-compatible API

    Exposes an API compatible with Elasticsearch.

  • Reference coding

    Index open-source repos so agents consult real definitions.

  • MCP integration

    Stdio MCP server connects to a running node.

Refy vs XERJ: questions

Is Refy or XERJ cheaper?

Both are open source. Compare current plans on each official site.

What is the difference between Refy and XERJ?

Refy is an open-source MCP server for developer tools. XERJ is an open-source MCP server for knowledge and memory.