Lune Research is an MCP (Model Context Protocol) server used to connect AI agents to top-tier computer science literature, enabling them to obtain evidence-based, trustworthy answers in their research work and avoid hallucinated facts and open web noise.
[Purpose] Lune is aimed at academic research scenarios, helping AI agents produce grounded rather than guessed results in every part of the research workflow, including literature search, academic writing, literature review, experimental design, and peer review. It is officially stated to be used by students and researchers at multiple institutions.
[Main Capabilities]
- Trusted knowledge sources: based on A/A* conference papers (such as NeurIPS, IEEE S&P, ACL); the official description states that workshop, findings, tutorial, short, abstract, or demo papers are not included, in order to reduce open web noise.
- Full-text semantic search: search the paper library in one go using natural language.
- Citation tracing: see what a paper builds on and what it leads to.
- Research best practices: provide validated research guidance when relevant to the research.
- Access the ideas of great researchers: search talks, interviews, and notes from the people who shaped the field.
- Workflow coverage: from the initial literature search to the final draft, every step is evidence-based. Official examples include finding foundational papers, locating research gaps, designing evaluation baselines, and writing related-work positioning sentences.
[How to Use] The official offering provides multiple ways to connect:
- Install as a plugin or connector into an existing AI application (no need to download a new app).
- Connect through a custom connector.
- Install and use Lune in the terminal via the command line.
- Use Lune on the web.
- For custom-designed agents: built-in Lune MCP.
[Applicable Scenarios] Aimed at students and researchers, it is suitable for research tasks such as literature review, evaluation design, and paper writing, and can also be embedded in custom agents.
[Limitations]
- The official description states that its knowledge sources are top-tier computer science papers, and it does not mention coverage of other disciplines.
- The free version is 10 requests per day; the Pro version is 300 requests per month/day and requires a team sharing plan and quota; higher versions are 600 requests per day. Beyond the quota, 300 additional requests can be purchased for $0.99, and the quota never expires.
- Specific platform compatibility, supported regions, data privacy details, and similar information are not yet provided.
