Academic Research Assistant MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Academic Research Assistant MCP Serverhelp me plan a systematic literature review"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Academic Research Assistant
An MIT-licensed toolkit for research planning, literature work, scientific reasoning, writing, figures, and academic documents. It contains 38 research skills, a dependency-free local MCP server, optional Python inspection/plotting tools, and an explicit-command revision runner.
Skills are methods that a host assistant reads and follows. The MCP server discovers and reads these methods; it does not search scholarly databases, generate research results, compile documents, or call models by itself. External research tools remain separately configured host capabilities.
Install from GitHub
Download the npm tarball from GitHub Releases, or clone the editable source:
git clone https://github.com/santanu-p/academic-research-assistant.git
cd academic-research-assistantThe release also includes a portable plugin ZIP and SHA-256 checksums. For local tarball installation, see installation.
Related MCP server: WebMCP Script
Try the local source
Requires Node.js 22 or later. No Node dependencies or account setup are needed.
node cli/ara.mjs --help
node cli/ara.mjs list
node cli/ara.mjs route "plan a systematic literature review"
node cli/ara.mjs verifyBefore npm publication, use the local source or the release tarball. After publication the equivalent entry point will be npx academic-research-assistant@3.0.0 --help. The npm name is proposed; availability and publication have not been verified.
Use with a host
Run exports from this project directory. Choose a new destination: existing folders are refused. Export/install copy the complete runtime bundle and print the next steps. They do not change host settings or install host applications.
Host/profile | Command |
Codex local marketplace |
|
Claude Code plugin |
|
Codex single entry skill |
|
Claude single entry skill |
|
Other assistant / local MCP |
|
The plugin profiles expose individual skills. The single-entry profiles use the root SKILL.md to route to nested methods; keep the entire directory together. Generic compatibility depends on the host supporting local files, Agent Skills, or stdio MCP. There is no claim that every CLI or app supports these interfaces.
See installation and host support for registration commands and the verification boundary.
Research coverage
The skill catalog lists all methods. Start with research-assistant, which routes by research stage. Every skill shares the research contract: preserve evidence, identify assumptions, distinguish completed work from proposals, and verify artifacts before claiming completion.
Coverage includes research questions, methods, literature searching and appraisal, systematic reviews, meta-analysis, quantitative and qualitative analysis, citations, research integrity, proposals and planning, scientific models and validation, manuscript preparation and revision, LaTeX production and repair, conversion, figures and tables, slides, posters, theses, professional documents, teaching materials, and public communication.
Executable tools
Python 3.10+ is optional and separate from the Node runtime.
Tool | Purpose | Additional libraries |
| Descriptive prose counts and optional exemplar comparison | None |
| Inspect delimited numerical tables and uncertainty notation | None |
| Inspect SVG, raster, or PDF artifacts | Pillow for raster; PyMuPDF for PDF |
| Plot comparisons, intervals, trends and Pareto frontiers | NumPy, Matplotlib |
| Run two reviewers and a reviser supplied as explicit command arrays | None; caller supplies commands |
python -X utf8 -B tools/prose_metrics.py draft.md --json
python -X utf8 -B tools/table_qa.py results.tsv --json
python -X utf8 -B examples/scientific_figures.py --out outputs/synthetic
python -X utf8 -B tools/revise.py --helpSee the scientific helper API and revision runner. Inspection reports are aids to review, not scientific or stylistic certificates. No plagiarism or AI-authorship detector is included.
Develop and package
npm run build
npm run validate
npm test
python -X utf8 -B -m unittest discover -s tests/python
npm packNo npm install is required for the Node runtime. Python tests involving optional libraries report skips when those libraries are absent. Developer tests, generated outputs, and private recovery material are excluded from the npm tarball. See development, release preparation, and source disposition.
Version 3 is newly authored public source. Supplied source collections and previous private integrations are not vendored. LICENSE covers this project's new code and text; third-party notices describe the boundary. Rewriting existing third-party material would not remove its license conditions.
This server cannot be deployed
Maintenance
Related MCP Connectors
Read-only MCP over the LivingMeta AI-in-Research corpus: 12,400 papers, gaps, priority agenda.
Read-only MCP server exposing a user ORANO library to their own AI agent.
Machine-native research commons for agent evidence, discovery, rooms, and bounded research quests.
Discover tasks and reproducible results. Public reading; invited, host-approved contributions.
Related MCP Servers
- AlicenseAqualityDmaintenanceExposes a verified tool registry (calculator, sandboxed file read, web fetch) over MCP stdio, enabling any MCP-capable client to reuse the same tools from the inspectable ReAct loop.3MIT
- AlicenseBqualityBmaintenanceEnables AI clients to connect locally over stdio, discover and read schemas for native page tools, and execute user-authorized operations supplied by websites and independent user scripts.13MIT
- AlicenseNot gradedqualityCmaintenanceEnables coding agents to search local repositories and retrieve bounded, cited source chunks, look up full indexed chunks, and inspect index statistics through read-only MCP tools over stdio.MIT
- AlicenseNot gradedqualityAmaintenanceA local stdio MCP server that exposes tools for running evidence-first research with source-linked, replayable reports, plus replay, citation-audit, and Agent Reach health checks. It preserves fetched page text, exact quote offsets, hashes, and run traces so a client can verify that quotes match retrieved sources, while withholding arbitrary shell or local-file access.1MIT