BenchCost

Inspiration

Research teams may understand the biology but still struggle to answer a basic planning question: what will this experiment actually cost? The inputs are scattered across protocols, vendor catalogs, lab inventory, spreadsheets, and assumptions that change as the experiment changes. BenchCost was built to turn that fragmented planning work into one explainable estimate before anyone enters the lab.

What it does

BenchCost brings methods, product prices, lab inventory, and planning assumptions into a single interactive workspace. The current Western blot prototype turns sample count, replicates, controls, target, transfer method, annual frequency, and owned equipment into an editable 11-step plan.

As inputs change, BenchCost recalculates:

• purchase cost and per-experiment cost
• per-sample, per-lane, and annual usage
• materials, equipment, and working time
• linked workflow effects and compatibility
• assumptions and unresolved inputs that still need a decision

Users can preview the full impact of a transfer-method change before applying it, inspect line-item assumptions, choose compatible primary and secondary antibodies, compare configurations, save drafts locally, and export JSON or CSV.

How we built it

We built BenchCost as a browser-based Next.js and React application with TypeScript and Tailwind CSS. A deterministic cost engine connects experimental inputs to the 11-step workflow, linked-change previews, cost breakdowns, equipment requirements, and scenario comparison. OpenAI Codex with GPT-5.6 Sol supported implementation, UX iteration, English localization, and desktop/mobile QA.

Challenges we ran into

The hardest part was not displaying a total. It was keeping every decision explainable while multiple parts of the experiment changed together. A transfer-method change can affect equipment, consumables, time, compatibility, and cost at once, so the interface had to show those linked effects before the user committed the change.

Accomplishments that we're proud of

We created a problem-led English planning experience with live recalculation, an editable 11-step workflow, linked transfer-method previews, antibody compatibility guidance, line-item cost assumptions, unresolved-input tracking, local draft saving, and JSON/CSV export.

What we learned

Experimental cost is a property of the entire workflow, not a single price list. The most useful estimate shows what changed, why it changed, and which inputs still need evidence.

What's next for BenchCost

BenchCost will collect and structure more than one million data points from across the web to calculate the cost of your specific experiment. That expanded evidence base will support more protocols, more products, richer compatibility checks, institution-specific inventory, and increasingly precise experiment-level estimates.

Built With

• codex
• gpt-5.6
• next.js
• openai
• react
• tailwind-css
• typescript