Arxiv MCP Server (blazickjp) vs Codescene MCP Server
Two MCP servers from our science & health ranking, compared on live data - updated nightly, never sponsored.
Bottom line · 2026-08-25
Arxiv MCP Server (blazickjp) records significantly more monthly package installs (84,878 vs 13,317); both are actively maintained, with commits inside the last month.
How they differ in kind
Arxiv MCP Server (blazickjp) focuses on: An MCP server for searching arXiv, downloading papers, reading bounded full text, retrieving original LaTeX by section, following citation graphs, and maintaining research alerts. Codescene MCP Server focuses on: The CodeScene MCP Server exposes CodeScene's Code Health analysis as local AI-friendly tools. Implementation languages differ (Python vs Rust).
What each one does
Arxiv MCP Server (blazickjp)
An MCP server for searching arXiv, downloading papers, reading bounded full text, retrieving original LaTeX by section, following citation graphs, and maintaining research alerts. Its 21 documented tools cover paper, latex, papers, abstract. It runs locally over stdio via the published package.
From the project's README.
Codescene MCP Server
The CodeScene MCP Server exposes CodeScene's Code Health analysis as local AI-friendly tools.
From the project's README.
14 signals, side by side
| Signal | Arxiv MCP Server (blazickjp) | Codescene MCP Server |
|---|---|---|
| Monthly installs | 84,878 | 13,317 |
| GitHub stars | 3,071 | 61 |
| Last commit | 2026-08-24 | 2026-08-18 |
| Releases · last 90 days | 8 | 10+ |
| In the registry since | Apr 2026 | Oct 2025 |
| Registry versions | 10 | 4 |
| Documented tools | 21 | not extracted - see README |
| Env vars extracted | none extracted - see README | none extracted - see README |
| Runs | local (stdio) | local (stdio) |
| Endpoint auth | local only | local only |
| Maintenance | actively maintained | actively maintained |
| License | Apache-2.0 | NOASSERTION |
| Language | Python | Rust |
| Category rank | #1 of 60 | #4 of 60 |
What their tools cover
Extracted from each project's README - the tools each server hands your agent. A missing list means the README doesn't document one, not that no tools exist.
Arxiv MCP Server (blazickjp) 21 tools documented
search_papers- Search arXiv by query, category, date, and sort orderget_abstract- Fetch metadata and an abstract by arXiv IDdownload_paper- Download and convert a paper to local Markdownlist_papers- List papers stored locallyread_paper- Read locally stored paper contentget_paper_outline- Paginated markdown heading outlineread_paper_section- Read one bounded markdown sectionsearch_paper_text- Bounded passage search in a paper
+13 more on the full page
Codescene MCP Server no tool list in README
Which one, for what
Derived from the signals above, not hands-on testing.
Pick Arxiv MCP Server (blazickjp) if…
- → you want the more widely installed option - 84,878 monthly installs vs 13,317
- → clear licensing matters - Apache-2.0 vs undeclared
Pick Codescene MCP Server if…
- → you prefer a Rust codebase to extend or audit
- → category standing - #4 of 60 maintained science & health servers
Who's behind them
Arxiv MCP Server (blazickjp) - blazickjp: 1 MCP server tracked, 1 maintained, 3,071 combined stars.
Codescene MCP Server - codescene-oss: 1 MCP server tracked, 1 maintained, 61 combined stars.
Not sold on either?
The next-ranked science & health servers we track:
Quick answers
Is Arxiv MCP Server (blazickjp) better than Codescene MCP Server?
Package installs favor Arxiv MCP Server (blazickjp): 84,878 monthly installs to 13,317. Codescene MCP Server still ranks #4 of 60 maintained science & health servers. Data as of 2026-08-25; we haven't hands-on tested either.
Can I use Arxiv MCP Server (blazickjp) and Codescene MCP Server together?
Yes - MCP clients accept multiple servers in one config, so you can enable both science & health servers side by side. If their tools overlap, keep the one whose toolset fits to keep your agent's tool list lean.
Keep exploring: best science & health servers · Python servers · Rust servers · all comparisons · every server
Methodology: package-usage signal = npm/PyPI installs (last month, platform APIs) · maintenance = commit recency + release cadence (GitHub) · tool lists and environment-variable names extracted from each project's README · endpoint auth from our own nightly probes. We haven't hand-tested these servers; everything here is data as of 2026-08-25.