Better posters, faster, with AI that respects the science.

Struqo is an AI-assisted authoring platform for scientific and medical conference posters. Researchers paste an abstract, upload figures, and receive an editable PowerPoint poster built on evidence-based communication principles, typically in under five minutes.

The problem we solve

Most conference posters fail on arrival. They are miniature manuscripts: dense paragraphs, tiny fonts, no focal point. Reviewers walk past. The result is wasted research effort, missed collaborations, and slower dissemination of findings that matter. Existing tools (PowerPoint, LaTeX, Canva) hand the author a blank canvas and no guidance on what actually communicates.

Our approach

Struqo combines two things: an evidence-based layout system (built on Mike Morrison's Better Poster format) and a purpose-built AI pipeline that rewrites dense academic prose into scannable, hallway-readable content, without changing the meaning.

Inside the AI pipeline

AI is used at four specific stages. Every stage is transparent, editable, and the author remains the final decision-maker.

1. Abstract parsing. A large language model extracts structured fields (title, authors, affiliations, background, methods, results, conclusion, references, funding, and disclosed conflicts) from unstructured abstract text.
2. Plain-language rewriting. Each section is rewritten in active voice, short sentences, and concise phrasing. Even specialist audiences prefer writing that is easy to follow because it saves time and improves comprehension. Clear writing is not simplification; it is more effective communication.
3. Key-finding synthesis. The model derives 3 to 4 distinct key findings from the results, plus a separate one-sentence key takeaway that acts as the poster's headline. The takeaway and section title are generated independently so they do not repeat each other.
4. Compliance review. For regulated contexts (such as industry-sponsored medical posters), the model checks for the presence of required sections (Important Safety Information, fair balance statements, references, disclosures) and flags anything missing before export.

Models, data, and privacy

Struqo uses commercial frontier large language models accessed through a secure API gateway. We do not train models on user content. Abstracts, figures, and drafts belong to the author. Drafts are stored in the user's account so they can return and edit; they can be deleted at any time. Figures are downscaled in the browser before upload to keep payloads light and reduce storage of sensitive imagery.

The design system

Every generated poster inherits a strict design system enforced at render time: a typographic scale (H1 roughly 3 to 4 times body), a 4/8/16/24/32/48 spacing rhythm, WCAG-AA text contrast, an accent-area cap so highlight colours never overwhelm the eye, and per-template layouts (billboard, IMRaD, modular grid, summary rail, twin-finding, visual-first). Authors pick a template and a palette; the system applies the rules invisibly.

Output

Struqo exports a fully editable .pptx file. Nothing is rasterised. Authors can open the file in PowerPoint or Keynote, adjust anything, and print at conference scale. This matters because posters almost always need a last-mile edit, such as a co-author correction, a late result, or a sponsor logo.

Who it's for

• Academic researchers preparing conference posters under time pressure
• Medical affairs and medical writing teams producing congress deliverables
• Clinical trial teams communicating results to mixed audiences
• Startups and biotech companies presenting at scientific meetings

What's coming next

• More evidence-based templates for different disciplines
• Audience-tailored rewriting (clinician, patient, policy-maker)
• Alignment with adjacent deliverables such as slide decks, manuscripts, and plain-language summaries
• AI-assisted scientific illustration and figure generation
• Team workspaces with co-author review and version history