Free tool · works now

Research project starter kit builder

Assemble connected plain-text research repositories with source, evidence, synthesis, and draft templates.

Research Project Workspace Generator

Generate a connected, multi-file research repository with guaranteed internal link resolution and empirical evidence tracing.

Zero Cloud Telemetry100% Resolved LinksObsidian & CommonMark Native
Title rendered across generated dashboard and manuscript headers.
Principal investigator or graduate student name.
Tailors methodology prompts and theoretical frames.
Ensures links resolve natively in your chosen editor.
The foundational empirical question embedded into the dashboard, abstract, and synthesis notes.
Active Workflow Modules:
Generated Files6
Internal Links23
Link Status100% Resolved
Broken Links0 Broken
00-Project-Dashboard.md
# Research Project Dashboard: Mechanistic Interpretability of Sparse Autoencoders
**Principal Investigator:** Dr. Jane Chen
**Field / Discipline:** COMPUTER SCIENCE
**Initiated Date:** 2026-11-19
**Status:** In Progress (Active Investigation)
## Core Research Question
> **How do polysemantic features resolve into monosemantic activations under sparse dictionary learning?**
## Project Architecture & Artifacts
This research kit provides a fully connected plain-text knowledge network. Every artifact links bidirectionally to its parent dashboard and evidence trail:
- 📚 **Source Literature:** [[01-Sources/Source-Reading-Notes|Primary Literature & Reading Notes]] — Catalog of inspected papers and citations.
- 🔬 **Evidence Matrix:** [[02-Evidence/Evidence-to-Claim-Matrix|Evidence-to-Claim Matrix]] — Empirical grounding connecting assertions to page citations.
- 💡 **Literature Synthesis:** [[03-Synthesis/Literature-Synthesis|Literature Synthesis & Contradiction Map]] — Thematic comparison across competing perspectives.
- ✍️ **Manuscript Draft:** [[04-Writing/Manuscript-Draft|Manuscript Working Draft]] — Sectioned manuscript referencing established claims.
- 🧭 **Decision Journal:** [[05-Meta/Research-Decision-Log|Methodological Decision Log]] — Immutable log of research choices and hypotheses.
## Working Hypotheses & Epistemic Boundaries
- [ ] **Hypothesis 1:** Primary mechanism accounts for observed empirical variances.
- [ ] **Hypothesis 2:** Counter-evidence from prior publications can be reconciled through scope refinement.
## Current Sprints & Milestones
- [ ] Complete initial reading of foundational texts in [[01-Sources/Source-Reading-Notes|Source Reading Notes]].
- [ ] Populate the first 5 claims in the [[02-Evidence/Evidence-to-Claim-Matrix|Evidence Matrix]].
- [ ] Draft methodology and experimental design in [[04-Writing/Manuscript-Draft|Manuscript Draft]].

The Architecture of Plain-Text Research Vaults: Link Integrity and Evidence Traceability

In academic research, investigative reporting, doctoral scholarship, and scientific policy analysis, researchers confront an escalating volume of primary literature, quantitative datasets, and methodological decisions. When starting a major project, many investigators resort to fragmented organizational habits: PDF highlights stored in proprietary reference managers, meeting notes buried in cloud documents, and working manuscript drafts disconnected from underlying citations.

This organizational fragmentation produces severe epistemic vulnerabilities: citations drift from their original context, empirical claims in drafts lack direct quotation anchors, and researchers waste dozens of hours re-reading papers they already summarized. Our Master Plan (§TOOL-076) establishes a rigorous, sovereign paradigm: the plain-text research starter kit with 100% internal link resolution.

1. The Five Essential Research Artifacts in Plain-Text Systems

A complete, resilient research repository requires five distinct interconnected artifacts, each serving a unique stage in the cycle of scientific inquiry:

2. The Mathematical Requirement of Internal Link Resolution

A foundational requirement mandated in our Master Plan (§TOOL-076) dictates that all internal sample links must resolve strictly within the generated package.

Many generic template packs provide boilerplate text containing hypothetical links (such as [[Smith-2024]] or [[Appendix-B]]) without actually providing the target notes. When a researcher imports these templates into Obsidian or Logseq, their vault graph immediately litters with broken links and dead ends.

Our generator algorithmically inspects every outgoing wikilink and relative file path across all generated templates. If an optional module (such as the Decision Journal) is toggled off, the engine automatically purges links pointing to that module across the dashboard and draft templates. When you import the generated package into your vault, every single link resolves to an existing file with 0 broken targets.

3. Methodological Pre-Registration and Decision Logging

In modern experimental science and psychology, the replication crisis highlighted the danger of p-hacking, retrospective hypothesis generation (HARKing), and subtle publication bias. When researchers encounter unexpected friction, they frequently alter parameter boundaries or sample exclusions without documenting the rationale.

The Methodological Decision Journal (05-Meta/Research-Decision-Log.md) provides an immutable audit trail within your knowledge vault. By logging critical methodological choices, parameter calibrations, and rejected alternative pathways in real time, you establish transparent provenance that withstands rigorous peer review.

4. Wikilinks vs CommonMark Hyperlinks: Choosing Your Vault Format

The starter kit builder supports two major link syntax standards:

5. Formula Injection Defense and Air-Gapped Security

Scientific research projects often involve sensitive intellectual property, unreleased algorithmic code, confidential medical trial data, or proprietary patent disclosures. Cloud-based research assistants upload your research questions and notes to hosted language model servers.

The Research Project Starter Kit Builder runs entirely client-side in your web browser. All file generation, link graph validation, and CSV manifest serializers execute within local JavaScript memory. No research hypotheses or project titles are ever transmitted over the network.

6. Provenance Architecture: Connecting BibTeX Keys to Atomic Quotes and Manuscript Paragraphs

The core vulnerability in traditional academic composition is the "citation leap"—the unverified chasm between a broad reference entry in a bibliography and a specific factual assertion made in the body of an article.

In conventional writing workflows, an author types a sentence such as: "Sub-linear scaling persists across distributed transformer topologies." The author attaches a general citation key (\cite{smith2025}). Two years later, during rigorous peer review, a referee requests the precise experimental conditions, baseline hyperparameter tuning, or page location where that scaling exponent was established. The author is forced to conduct an agonizing manual search through a 40-page PDF to rediscover the evidence.

Our Research Starter Kit resolves this cognitive friction through an unbroken three-tier provenance trail:

  1. Tier 1: Atomic Source Notes: Foundational papers are summarized with verbatim extracts and exact page indicators (e.g., p. 114, paragraph 3) in 01-Sources/Source-Reading-Notes.md.
  2. Tier 2: The Empirical Evidence Matrix: Each factual assertion receives a deterministic identifier (e.g., Claim C-01) in 02-Evidence/Evidence-to-Claim-Matrix.md, connecting the extracted quotation directly to the thesis assertion.
  3. Tier 3: The Working Manuscript: In 04-Writing/Manuscript-Draft.md, draft paragraphs embed direct wikilinks or relative references to the Evidence Matrix rather than loose mental impressions.

7. Multi-Investigator Collaboration in Git-Versioned Markdown Repositories

When research teams collaborate using commercial office suites or cloud word processors, track changes quickly devolve into versioning conflicts: Manuscript_v2_final_FINAL_edits_revised.docx. Furthermore, proprietary binary formats make line-by-line attribution and automated linting impossible.

By grounding your investigation in plain-text Markdown files generated by this kit, your research repository gains immediate access to the entire modern software engineering ecosystem:

8. Peer Review Preparation: Exporting Evidence Bundles as Supplementary Material

In high-stakes academic publishing and regulatory policy filings, transparency is the ultimate differentiator. Reviewers and committee chairs increasingly demand open science transparency, requesting access to underlying literature audits and empirical calculation logs.

Because our starter kit compiles your complete investigation into a cohesive, linked package, preparing supplementary materials requires zero manual formatting. You can export the Evidence-to-Claim Matrix and Literature Synthesis directly into your replication package or Zenodo preprint repository. Reviewers can verify the exact quotation provenance for every claim in your manuscript with a single click, dramatically accelerating the path to peer acceptance.

9. Resilient Open Science: Escaping Proprietary Reference Manager Lock-In

Over the past two decades, academic researchers have repeatedly suffered when commercial publishers acquired and closed independent reference management software. Proprietary databases lock your marginalia, tags, and synthesis notes behind subscription paywalls or proprietary file formats.

Plain-text Markdown files and standard BibTeX keys are open international standards that will remain readable and parsable for centuries. By establishing your research project on a sovereign, local-first starter kit, you guarantee that your intellectual discoveries remain permanently accessible, independent of corporate acquisitions, platform pivots, or software licensing fees.

Questions

Why does this tool strictly guarantee 100% internal link resolution?

Starting a complex research project with broken or mismatched file templates creates immediate friction in your knowledge graph. If your manuscript draft links to an evidence file with a different filename or missing path, your bi-directional graph shatters. Our generator algorithmically verifies that 100% of internal links between the dashboard, sources, evidence matrix, synthesis notes, and draft point strictly to files that exist within your package.

How does the Evidence-to-Claim Matrix prevent academic hallucinations?

When authoring long-form manuscripts, researchers often write assertions from memory that diverge from the underlying literature. The Evidence-to-Claim Matrix establishes an empirical bridge: every factual claim in your draft links directly to an extracted quotation, exact page number, and bibtex citation key in your source reading notes.

Can I use this research kit outside of Obsidian?

Yes. The starter kit builder supports both Obsidian [[WikiLinks]] and RFC 3986 CommonMark relative file links (./01-Sources/Source-Reading-Notes.md). The generated Markdown files can be opened in Visual Studio Code, Neovim, Zettlr, Logseq, or pushed directly to private GitHub repositories.

What is the purpose of the Methodological Decision Journal?

In empirical research, investigators frequently encounter unexpected roadblocks and alter parameters or hypothesis scopes. Without real-time documentation, researchers fall victim to hindsight bias during peer review. The Decision Journal provides an immutable log of methodological choices, parameters tested, and rejected alternatives.

Are my confidential research questions, hypotheses, or manuscripts sent to external servers?

Never. The entire starter kit builder, file linker, link validation engine, and export serializers run 100% locally in your browser memory. No proprietary hypotheses or project details are ever transmitted over the network.

Launch Your Research with Confidence on Oyma

Construct sovereign, fully connected plain-text research repositories without sacrificing your privacy to cloud surveillance.

tools research project starter kit builder downloadDownload for Mac

No account needed. Intel Mac build