From vague idea to conference poster: an AI co-author pipeline

What it does ResearchStudio is a pair of skill suites for AI coding agents that bookend the actual writing. The Idea module takes a vague research direction and distills it into a reviewer-defensible idea card, backed by an empirical taxonomy drawn from thousands of ICLR, ICML, and NeurIPS submissions. The Reel module ingests a finished paper PDF and produces the post-publication collateral: a print-ready poster, narrated walkthrough video, bilingual blog post, and an interactive reel viewer. Everything runs inside Claude Code or OpenAI Codex, not as a standalone application.

The interesting bit Most AI writing tools focus on the draft; ResearchStudio treats the “before” and “after” as equally ripe for automation. The ideation component is notably methodical—it tries to ground creativity in historical acceptance patterns rather than generic brainstorming, which is either reassuring or deeply cynical depending on your view of peer review.

Key highlights

• End-to-end coverage: pre-paper ideation (Idea) and post-paper dissemination (Reel)
• Idea generation grounded in an empirical taxonomy derived from major ML conference submissions
• Reel generates posters, narrated videos, bilingual blogs, and an interactive reel from a single PDF
• Ships as installable skills for Claude Code and OpenAI Codex, not a monolithic app
• MIT licensed, with two accompanying arXiv papers

Caveats

• The README cites model versions (claude-opus-4.6, gpt-5.5) and 2026 release dates that do not exist yet, so the project appears forward-looking or speculative
• Functional demos for both modules are marked “Coming Soon”
• Because these are agentic skills rather than a packaged application, setup requires familiarity with Claude Code or Codex environments

Verdict Researchers who already live inside Claude Code or Codex and want to automate literature packaging should try it; if you need a polished GUI or are wary of AI-generated research directions, look elsewhere.