Interpret any graph, chart, or research figure with Crade
A traffic spike chart, a research figure with three axes, an unfamiliar visualisation in a paper. Crade reads it on your screen and tells you what it actually shows and what to take away.
| Source | Visitors | Conv. |
|---|---|---|
| Product Hunt | 4,821 | 3.2% |
| Google (organic) | 3,456 | 4.8% |
| Twitter / X | 2,103 | 1.1% |
| Direct | 1,892 | 5.6% |
| Hacker News | 1,240 | 0.8% |
What does this graph show?
Your traffic over 30 days: • Spike on Mar 5 → Product Hunt launch (+340%) • Dropped 60% by Mar 8 → visitors didn't stick • Bounce rate up 5.4%, session time down 18% → Focus on retention: add onboarding flow or email capture
Charts and graphs are designed to communicate, but they often fail. Bad axis choices, unfamiliar visualisation types, missing context, dense scientific notation. Reading them takes effort even when they are well-designed. Crade reads the chart on your screen and gives you the takeaway: what the data shows, what is unusual, what to act on if anything.
What you put on your screen
- A chart in a dashboard (analytics, finance, monitoring)
- A figure in a research paper
- A graph in a news article or report
- A chart in someone else's presentation
- Your own visualisation that you are not sure communicates clearly
What you say to Crade
Or specific: "What is the spike on March 5?", "Is the downward trend significant?", "What is this figure trying to communicate?".
Step-by-step: how to do this in Crade
Open the chart at full size
Make sure the axes, labels, and data points are readable.
Click the Crade icon
Expand Crade.
Ask the question
Plain question. "What is this showing?", "What changed?", "Why is the line going down?".
Read the interpretation
Crade describes the chart in plain language plus the takeaway. For unusual visualisations, Crade explains the format first, then the data.
Ask follow-ups
"What might cause that spike?", "How does this compare to last month?", "What is the unit on the Y axis?".
Use the takeaway
Make the call, write the report, ask the next question. The chart did its job because Crade did the interpretation.
What you get back
A plain-language interpretation: what the chart shows (X over Y), the notable patterns (trends, outliers, inflection points), the likely cause if obvious, and the takeaway. For research figures, Crade also explains the format if it is unusual (violin plots, Sankey diagrams, etc.).
Tips for better chart interpretation
- Tell Crade what you are looking for. "I am trying to decide if conversion is improving" steers the interpretation toward that question.
- If the chart has multiple series, ask Crade to focus on one. "Just the blue line" is faster than asking about all of them.
- For unfamiliar chart types, ask Crade to explain the format first: "What kind of chart is this?".
- Verify any number Crade quotes from the chart. AI is good at narrative, occasionally fuzzy on exact values.
- For research papers, ask for the methodology too: "What is the statistical test behind this figure?".
Free vs Pro vs Premium
Chart interpretation works on every plan. Higher tiers give you higher daily usage.
Frequently asked questions
How accurate is Crade at reading chart values?
Good at narrative ("X is trending up"), less precise on exact numbers ("X is at 47.3"). For specific values, hover over the chart in the source tool to get the exact figure rather than relying on Crade's eyeball read.
Can Crade analyse interactive charts?
Crade reads the static visible state. For hover tooltips or zoomable charts, the values are what was visible at the moment of the prompt. Re-prompt after changing the view if needed.
Does Crade understand scientific notation and uncommon plot types?
Yes for common scientific formats (log-scale axes, error bars, violin plots, heatmaps). For very specialised or custom visualisations, ask Crade if it is confident. It will tell you.
Can Crade tell me if a chart is misleading?
Yes, and often helpfully. "Is this Y-axis truncated in a way that exaggerates the trend?" is a great use of Crade. Standard visual tricks (cut axes, misleading scales, cherry-picked time ranges) are easy for Crade to spot.
Should I trust Crade's interpretation for high-stakes decisions?
Use it as a first read. For decisions that matter (board meetings, investment calls, scientific publication), back the interpretation with your own check of the underlying data.
The whole loop in one sentence
Chart on screen, one prompt, the takeaway back. From "what am I looking at" to "now I know what to do" in thirty seconds.
A screenshot of a slide. A photo of a printed report. A frame from a YouTube video. Crade reads the table on your screen and gives you the rows in CSV format, ready to paste into a spreadsheet.
Three plans, three products, three job offers, three Airbnbs. Open them in separate tabs or windows. Crade reads them all and tells you which one fits your situation best, and why.
A 30-page vendor agreement, a 50-page report, a research paper. Crade reads what is on your screen and gives you the structured summary: what it is about, the key points, what to act on.